Header

Behavior

Scope

Body mismatch on replay

When to use

Without an Idempotency-Key

Idempotency

People discovery, contact enrichment, and personalized outreach. One API for your apps and agents.

 Welcome to Clodo Find the right people, enrich their contact details, and turn research into personalized outreach — all through a single API. Get started Create an account, get your API key, and make your first request. APIs Search for people with natural language, resolve their professional identities, and reach them from your own connected inbox. Build a workflow Start with a people search API use case guide-use-cases , then connect the endpoints your application needs: - GTM lead generation guide-gtm — find B2B prospects, qualify decision-makers, enrich contact details, and prepare sales outreach. - Recruiting and candidate sourcing guide-recruiting — turn job requirements into a candidate search, review professional evidence, and draft recruiter messages. - Expert discovery guide-experts — identify practitioners for industry interviews, customer research, and advisory projects. Build with your AI tools Every page is available as Markdown. Use Copy page to bring the documentation into your editor, or Open in ChatGPT and Open in Claude to ask questions about a specific page. The documentation index https://docs.clodo.ai/llms.txt gives agents a list of all available pages. Download the full documentation https://docs.clodo.ai/llms-full.txt for a single text reference. Developer resources - Authentication authentication — API keys and request headers. - Webhook events webhook-events — receive search and enrichment results. - Async polling async-polling — follow a job through completion. - Handling errors errors — understand failures and retry safely. 

Welcome to Clodo

Create a Clodo account and generate your API key.

 Get an API Key Create an account 1. Open Get Started {{signup}} and choose API pay as you go . 2. Complete signup and checkout, then open the API Console {{console}} . 3. Check your balance in Credits before making a paid request. See Credits & Pricing credits for endpoint costs. Get your API key 1. Open API Console → API Keys {{console}}?tab=keys . 2. Click Mint new key and complete the key creation form. 3. Copy the raw key immediately. Allow up to 60 seconds before first use. Your API key is shown only once. Store it in a secret manager or server environment variable. Keep it out of browser code, public repositories, and shared screenshots. If it is lost or leaked, revoke it in the console and create a new key. Authenticate your requests Pass the raw key in the x-api-key header. Do not add a Bearer prefix or place the key in the URL. http x-api-key: ck_live_<your-secret The API base URL is https://api-public.clodo.ai/api/public/v1/ . Include the trailing slash on endpoint paths. Make your first API call Follow the Quickstart quickstart for a complete request, or choose an endpoint in the API Reference. For People Search, Deep Search, and Phone Enrichment, create a webhook signing secret webhook-secrets before making your first call. These endpoints require a public HTTPS webhook URL, even when you also use polling. 

Get an API Key

Go from your API key to your first enrichment response.

 Quickstart Three steps from zero to your first response: mint an API key, send a request, read the result. 1. Mint an API key Open API Console → API Keys https://clodo.ai/api-console?tab=keys and click Mint new key . Copy the raw secret immediately. We display it once and never store it in plaintext after that. Keys look like ck_live_<base64 . 2. Send your first request bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/email/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "domain": "stripe.com" }' A successful response: json { "email": "patrick@stripe.com" } Response headers include X-Request-Id a UUID per request . 3. Same call from Python python import requests response = requests.post "https://api-public.clodo.ai/api/public/v1/enrich/email/", headers={"x-api-key": "ck_live_..."}, json={ "first_name": "Patrick", "last_name": "Collison", "domain": "stripe.com", }, timeout=30, response.raise_for_status print response.json "email" 3b. Same call from Node javascript const response = await fetch "https://api-public.clodo.ai/api/public/v1/enrich/email/", { method: "POST", headers: { "x-api-key": "ck_live_...", "content-type": "application/json", }, body: JSON.stringify { first_name: "Patrick", last_name: "Collison", domain: "stripe.com", } , }, ; if !response.ok { throw new Error HTTP ${response.status}: ${await response.text } ; } const { email } = await response.json ; console.log email ; Beyond enrichment: full outbound The API also covers outreach itself — find people with Deep Search endpoint-deep-search , enrich their emails, then draft, review, and send from your own inbox with Outreach Emails endpoint-outreach-emails : bash Draft an email 202 - poll - review - approve - it sends curl -X POST https://api-public.clodo.ai/api/public/v1/emails/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "to": {"email": "jackie@acme.com", "first_name": "Jackie", "company_name": "Acme"}, "instructions": "Casual note asking if she is open to a chat about our founding PM role." }' With the default review: "required" , nothing sends without your explicit approval. No webhook receiver is required — polling is first-class for these endpoints. What to read next - Authentication authentication for the full key model and what headers we accept. - API Base URL base-url for the endpoint surface and which host to point at. - Per-endpoint reference for input shapes, output shapes, and credit costs. - People Search, Deep Search, and Phone Enrichment return 202 Accepted and POST results to a webhook URL you specify. Webhook Signing Secrets webhook-secrets and Webhook Signature Verification webhook-signing cover wiring a receiver. Async Polling async-polling is the fallback when your receiver is unavailable. Costs in this example 5 credits per call. See Credit Semantics credits for the full table. 

Make your first API call

Understand how Clodo credits work, what each endpoint costs, and how to manage your balance.

 Credit Semantics Per-endpoint cost Endpoint Cost Charge model --- --- --- Email Enrichment 5 credits Per call Phone Enrichment 50 credits on hit, 0 on miss Per hit Professional URL Enrichment 3 credits Per call Professional Profile Enrichment 2 credits Per call People Search 1 credit per lead returned Variable per result Deep Search 200 credits mode: "standard" or 400 credits mode: "extended" Fixed per run Agentic Search endpoint-agentic-search Variable, capped by max_credits 300–2,000; default 500 Reserve the cap, settle the final charge, release or refund the remainder Outreach Emails 2 credits per drafted email Per successful draft Polling endpoints GET /v1/.../{job_id}/ , GET /v1/emails/... consume no credits. For Outreach Emails, drafting is the only billable event — reviewing, editing, approving, cancelling, and sending are free, and a failed draft is not charged. Cancelling an already-drafted email does not refund the draft. Charge lifecycle Agentic Search charges 100 credits for a completed search that performs work but returns no qualifying leads. Internal failures and cancellation before paid work starts are not charged. Partial deliveries remain billable within your cap. See Agentic Search endpoint-agentic-search for the full budget and cancellation behavior. Each call creates an audit row that ends in one of these states visible in the Usage tab of the Console : charge_status Customer pill Meaning --- --- --- pending Pending In flight async endpoints, before settlement . charged Completed Charged the committed amount. zero_result Completed Variable-charge endpoint returned 0 results, no charge. released Completed Validation reject or upstream error, no charge. lost_ states are operational anomalies. Topping up Use the Credits tab in the Console. See also - Per-endpoint reference pages for input shapes and full status mapping. 

Credits & Pricing

Authenticate your requests and manage your API keys.

 Authentication Every request needs the x-api-key header. x-api-key: ck_live_<your-secret Keys are minted in API Console → API Keys https://clodo.ai/api-console?tab=keys . Minting an API key 1. Open the API Keys tab. 2. Click Mint new key . 3. Copy the raw secret. It is shown once. The table shows the key prefix and last four characters after creation. Allow up to 60 seconds after minting before first use. Revoking a key Click the trash icon next to a key. Revocation is immediate. In-flight requests with that key fail at the next request boundary with 403 Forbidden or 401 unauthorized . Other keys on the account keep working. Header rules - Header name is x-api-key . - Send the raw secret value. No Bearer prefix, no base64 wrapping. - Do not put the key in the URL or request body. Errors - 403 Forbidden — x-api-key header missing, revoked, or not recognized. Returned by the edge with body {"message":"Forbidden"} . - 402 insufficient_credits — balance hit zero. Top up in the Credits tab. - 403 permission_denied — key is valid but the account is not eligible for the public API. See Error Envelope errors for the full schema and code list. 

Authentication

The base URL, endpoint paths, and conventions for the Clodo API.

 API Base URL https://api-public.clodo.ai/api/public/v1/ Full URLs: Endpoint URL --- --- Email enrichment https://api-public.clodo.ai/api/public/v1/enrich/email/ People Search https://api-public.clodo.ai/api/public/v1/search/ Deep Search https://api-public.clodo.ai/api/public/v1/deep-search/ Agentic Search https://api-public.clodo.ai/api/public/v1/agentic-search/ Phone enrichment https://api-public.clodo.ai/api/public/v1/enrich/phone/ Professional URL https://api-public.clodo.ai/api/public/v1/enrich/professional-url/ Professional profile https://api-public.clodo.ai/api/public/v1/enrich/professional-profile/ Outreach Emails https://api-public.clodo.ai/api/public/v1/emails/ Sequences https://api-public.clodo.ai/api/public/v1/sequences/ Sending Accounts https://api-public.clodo.ai/api/public/v1/email/accounts/ Trailing slashes Always send the trailing slash. Endpoint paths are registered at the edge with a trailing slash /enrich/email/ , not /enrich/email . TLS HTTPS only. Plain HTTP is rejected at the edge. Versioning Version lives in the path: /api/public/v1/ . Breaking changes ship as /v2/ . /v1/ runs for at least 12 months after /v2/ ships, with Sunset response headers starting 6 months out. 

API Base URL

The canonical agentic people search endpoint. Describe the people you want to find in as much detail as you like—their role, background, expertise, company, or specific experiences. Clodo searches and researches potential matches, returning people with evidence explaining why they fit your criteria.

 Agentic Search Clodo's canonical agentic people search endpoint. Describe your target persona in plain language and get as specific as you want: their current role, past experience, technical expertise, location, the kind of company they work at, or something particular they have done. Combine criteria, explain what matters most, and include exclusions or examples to clarify who you are looking for. For example: “Find engineering leaders at US robotics companies who have personally worked on autonomous warehouse systems. Prioritize people who previously took a robotics product from prototype to production, and exclude consultants and recruiting firms.” Clodo searches people data and the open web, researches potential matches against your criteria, and returns people with fit summaries, proof points, and source evidence. The query field supports up to 2,000 characters. Use case Use Agentic Search to find people whose fit depends on a specific combination of experience, expertise, and accomplishments. Describe your ideal persona in detail, including what matters most and who to exclude. The agent researches potential matches and returns evidence explaining why each person fits. - Find B2B leads for a specific customer profile. Look for engineering leaders at robotics companies who have taken autonomous warehouse systems into production, or finance executives with experience expanding a business internationally. Explain the experience that makes someone relevant to your product. - Source candidates for hard-to-fill roles. Search for founding engineers who have built developer tools, operations leaders who have scaled manufacturing, or researchers with both academic and industry experience. Spell out the background your hiring team needs beyond a job title. - Find experts for interviews, advisory work, and research. Identify practitioners who have implemented a particular technology or solved a specific operational problem. Use the returned fit summaries, proof points, and source evidence to assess whom to contact. For complete workflows, see Recruiting & Candidate Sourcing guide-recruiting , Expert Discovery guide-experts , and GTM Lead Generation guide-gtm . Availability Agentic Search is enabled per account. Once enrolled, all your API keys can use it, including newly created keys. Accounts that are not enrolled receive 404 . Contact support mailto:hello@clodo.ai to request access. Credit budget Set max_credits to control the maximum charge for a job: 300–2,000 credits , with a default of 500 . The cap is reserved when the job is accepted. At completion, only the final charge is retained; the unused reservation is released or refunded. The final charge depends on the research performed and leads delivered, and never exceeds your cap. A completed search that performs work but finds no qualifying leads costs 100 credits . Internal failures and cancellation before paid work starts are not charged. Partial deliveries remain billable within the same cap. A webhook delivery failure does not undo the charge; retrieve the result by polling. Request body Field Type Required Description --- --- --- --- query string Yes Describe the people you want to find in plain language, up to 2,000 characters. Be as specific as you like about their role, background, expertise, company, location, or experiences. Include priorities, exclusions, and any evidence you want checked. Must not be empty. target_results integer No Desired number of leads, from 1 to 50. Default: 50 . This is a target, not a guaranteed result count. max_credits integer No Maximum charge, from 300 to 2,000 credits. Default: 500 . webhook_url string Yes Public HTTPS receiver URL, up to 2,048 characters. Must resolve to a public IP address. Create an active webhook signing secret webhook-secrets before submitting a request. Without one, the endpoint returns 409 . Request examples bash curl --request POST 'https://api-public.clodo.ai/api/public/v1/agentic-search/' \ --header "x-api-key: $CLODO_API_KEY" \ --header 'Content-Type: application/json' \ --header 'Idempotency-Key: agentic-search-example-001' \ --data '{ "query": "Find engineering leaders at US robotics companies with evidence of deploying autonomous warehouse systems.", "target_results": 25, "max_credits": 500, "webhook_url": "https://your-app.example/webhooks/clodo" }' python import os import requests response = requests.post "https://api-public.clodo.ai/api/public/v1/agentic-search/", headers={ "x-api-key": os.environ "CLODO_API_KEY" , "Idempotency-Key": "agentic-search-example-001", }, json={ "query": "Find engineering leaders at US robotics companies with evidence of deploying autonomous warehouse systems.", "target_results": 25, "max_credits": 500, "webhook_url": "https://your-app.example/webhooks/clodo", }, timeout=30, response.raise_for_status job = response.json print job "id" javascript const response = await fetch "https://api-public.clodo.ai/api/public/v1/agentic-search/", { method: "POST", headers: { "x-api-key": process.env.CLODO_API_KEY, "Content-Type": "application/json", "Idempotency-Key": "agentic-search-example-001", }, body: JSON.stringify { query: "Find engineering leaders at US robotics companies with evidence of deploying autonomous warehouse systems.", target_results: 25, max_credits: 500, webhook_url: "https://your-app.example/webhooks/clodo", } , } ; if !response.ok throw new Error Clodo returned ${response.status}: ${await response.text } ; const job = await response.json ; console.log job.id ; Run these examples on your server to keep your API key private. Replace the example webhook URL with your receiver and use a new idempotency key for each new search. Accepted response HTTP 202 Accepted : json { "id": "dj_a1b2c3d4e5f6478890abcdef123456789", "status": "pending", "created_at": "2026-09-07T12:00:00Z" } Results and webhook events The agentic_search.completed event contains the final result, including successful searches with zero leads: json { "id": "dj_a1b2c3d4e5f6478890abcdef123456789", "event_type": "agentic_search.completed", "partial": false, "completion_reason": "", "total_returned": 1, "results": { "full_name": "Alex Morgan", "current_job_title": "VP of Engineering", "headline": "Building warehouse robotics", "location": "Boston, Massachusetts, United States", "professional_url": "https://www.linkedin.com/in/example-alex-morgan", "company": { "name": "Example Robotics", "domain": "robotics.example", "industry": "Robotics", "employee_count_range": "51-200" }, "experience": {"company": "Example Robotics", "title": "VP of Engineering", "date_range": "2022-present"} , "fit_summary": "Leads engineering for a US company deploying autonomous warehouse systems.", "proof_points": "Company team page identifies Alex as VP of Engineering." , "evidence": {"headline": "Leadership team", "url": "https://robotics.example/team"} , "tier": "perfect" } } The response above is illustrative. Individual person fields may be empty or omitted when unavailable. Evidence entries can include headline , description , and url . Source URLs are only included after retrieval during the job. Field Meaning --- --- results Delivered people, with professional details, company, experience, fit summary, proof points, evidence, and a fit tier when available. total_returned Number of people delivered. A zero-result search returns 0 and an empty results array. partial true when the job stopped early and returned the results saved so far. completion_reason Empty on normal completion; budget_exhausted , deadline_exceeded , or user_cancelled when stopped early. An internal failure emits agentic_search.failed : json { "id": "dj_a1b2c3d4e5f6478890abcdef123456789", "event_type": "agentic_search.failed", "error": {"code": "error", "message": "Internal error — you were not charged."} } Verify webhook signatures webhook-signing before processing results. See Webhook Events webhook-events for additional failure codes. Polling Use GET /api/public/v1/agentic-search/{job_id}/ with your API key. Poll every 30–60 seconds until the status is completed , failed , or cancelled . The polling envelope async-polling includes status , timestamps, tier: "agentic_search" , leads_returned , result , and error . On completion, result contains the same result fields as the webhook, without id or event_type . Polling consumes no credits. Cancellation Send POST /api/public/v1/agentic-search/{job_id}/cancel/ with your API key. No request body is required. It returns 202 with the job's current id , status , and created_at . - A pending job cancels immediately and releases its entire credit reservation. - A running job stops at the next safe point and completes with partial: true , completion_reason: "user_cancelled" , and any saved results. Work already performed remains billable within your cap. - Cancelling an already terminal job returns its current state. It does not restart the job or refund completed work. Poll after requesting cancellation to confirm the terminal state. The cancellation request is idempotent. Idempotency The optional Idempotency-Key header is scoped to your API key. Repeating the same key and validated request body returns the existing job, without another dispatch or reservation. Changing the body—including webhook_url —while reusing the key returns 409 . Use a fresh key for every new logical request, including requests to other endpoints. Rate limits Operation Sustained requests per minute Burst --- --- --- Create search 2 2 Poll search 3,600 100 Cancel search 60 10 At most 2 Agentic Search jobs per API key and 3 per account can be pending or running. New requests beyond these concurrency limits return 429 . Replays of existing idempotent requests do not create a new job. Status codes HTTP Meaning --- --- 202 Search accepted, existing job replayed, or cancellation acknowledged. 400 Invalid query, target, cap, or webhook URL. 402 Billing cannot authorize the requested credit reservation. 403 API key is missing, invalid, revoked, or the account is ineligible for API access. 404 Account is not enrolled for Agentic Search, or the requested job is unavailable to your account. 409 No active webhook signing secret, or an idempotency key was reused with a different body. 429 Request rate or in-flight job limit exceeded. 500 Internal admission or dispatch failure; the reservation is released or refunded. For other platform errors and response formats, see Handling Errors errors . 

Agentic Search

Find sales leads, source candidates, and discover industry experts with a natural-language people search. Describe your target audience and retrieve matching professional profiles.

 People Search POST /api/public/v1/search/ — async returns 202 Accepted . 1 credit per lead returned. Light people search. Evaluates a natural-language query and returns matching leads. Wall time ~1 minute. Results delivered via signed webhook POST. Polling fallback at GET /api/public/v1/search/{job_id}/ . To fetch more non-overlapping leads for the same prompt, pass a completed job id back as search_id . Use case Use People Search to build a list of people from a natural-language description of your target audience. Search by professional criteria such as role, seniority, industry, company, and location, then retrieve matching profiles for your sales, recruiting, or research workflow. - Build prospect lists for outbound sales. Find heads of marketing at SaaS companies in London or procurement leaders at US manufacturers. Use their role and company details to organize leads before enriching contact information or preparing outreach. - Create candidate pools for recruiting. Source software engineers in Toronto, finance directors in New York, or product managers at fintech companies. Start with the professional profile you need and review the matches for your open role. - Discover experts and professional communities. Find renewable-energy executives for an industry interview series, healthcare operations leaders for customer discovery, or design leaders for an event. Search for the roles and sectors relevant to your project. When you need more people, continue a completed search with search_id to retrieve another non-overlapping page of results. Request Field Type Required Notes --- --- --- --- query string yes for new searches Natural-language sourcing prompt. ≤500 chars. Omit when continuing with search_id . search_id string no dj_... id from a previous completed People Search. Continues that search and returns the next non-overlapping page. max_results int yes min 25 and max 100. Upper bound on lead count. webhook_url string yes HTTPS URL. Must resolve to a public IP. Pre-condition: an active webhook signing secret. See Webhook Signing Secrets webhook-secrets . Idempotency-Key header is supported. Same key returns the same job's current state. Response 202 json { "id": "dj_a1b2c3d4...", "status": "pending", "created_at": "2026-05-01T12:00:00Z" } Use the returned id as search_id to continue pagination after the job completes. Webhook events search.completed : json { "id": "dj_a1b2c3d4...", "event_type": "search.completed", "results": { "first_name": "Patrick", "last_name": "Collison", "full_name": "Patrick Collison", "current_job_title": "CEO", "headline": "CEO at Stripe", "location": "San Francisco, California, United States", "professional_url": "linkedin.com/in/patrickcollison", "company": { "name": "Stripe", "domain": "stripe.com", "professional_url": "linkedin.com/company/stripe", "industry": "Financial Services", "employee_count": 8000 } } , "total_returned": 1 } Person and company professional_url values come back in bare-host form no scheme, no www. , no trailing slash . Company professional_url is optional and may be null . search.failed : json { "id": "dj_a1b2c3d4...", "event_type": "search.failed", "error": {"code": "error", "message": "People search failed; please retry."} } error.code error.message --- --- invalid_input Invalid search query. error People search failed; please retry. A 0-result outcome ships as search.completed with total_returned: 0 . Billing Charged per lead returned. max_results is the upper bound on credits. total_returned is the actual charge. A 0-result run charges nothing. Example bash curl -X POST https://api-public.clodo.ai/api/public/v1/search/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "query": "VPs of Engineering at Series B SaaS in NYC", "max_results": 50, "webhook_url": "https://yourapp.com/webhooks/clodo" }' Continue the same search: bash curl -X POST https://api-public.clodo.ai/api/public/v1/search/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "search_id": "dj_a1b2c3d4...", "max_results": 100, "webhook_url": "https://yourapp.com/webhooks/clodo" }' Rate limit POST /api/public/v1/search/ : Limit Value --- --- Sustained rate 6 requests / minute Burst 2 requests Polling GET /api/public/v1/search/{job_id}/ : Limit Value --- --- Sustained rate 3,600 requests / minute Burst 100 requests Exceeding any returns 429 . See also - GTM Lead Generation guide-gtm - Recruiting & Candidate Sourcing guide-recruiting - Expert Discovery guide-experts - Webhook Signing Secrets webhook-secrets - Webhook Signature Verification webhook-signing - Async Polling async-polling - Idempotency idempotency 

People Search

Find and qualify sales prospects, candidates, and subject-matter experts against detailed criteria. Turn a targeted people search into a researched shortlist with professional context and fit scores.

 Deep Search POST /api/public/v1/deep-search/ — async returns 202 Accepted . Fixed price per run: - mode: "standard" — 200 credits, ~25 qualified leads - mode: "extended" — 400 credits, ~100 qualified leads Wall time ~10-15 minutes. Results delivered via signed webhook POST. Polling fallback at GET /api/public/v1/deep-search/{job_id}/ . Use case Use Deep Search to find and qualify people against a focused set of criteria. It returns a researched shortlist with professional details, work experience, and fit scores, giving your team more context for deciding whom to approach. - Qualify sales leads before outreach. Search for decision-makers at companies in your target market—for example, technology leaders at regional banks or founders of climate-tech businesses in Germany. Review the returned company context and fit scores to prioritize prospects. - Build recruiting shortlists. Find candidates whose roles, industry backgrounds, and career experience align with a position. For a healthcare product leadership role, describe the relevant product and sector experience, then use the returned work history to guide your review. - Identify subject-matter experts for a project. Look for supply-chain leaders in manufacturing, executives in renewable energy, or operators in financial services. Compare their professional backgrounds when selecting potential interviewees, advisors, or research participants. Choose standard or extended mode based on the size of the shortlist you want. Each mode has a fixed price per successful run, making it straightforward to budget recurring prospecting, candidate sourcing, and expert discovery workflows. Request Field Type Required Notes --- --- --- --- query string yes Natural-language query. ≤500 chars. mode string no "standard" default or "extended" . webhook_url string yes HTTPS URL. Must resolve to a public IP. Pre-condition: an active webhook signing secret. See Webhook Signing Secrets webhook-secrets . Idempotency-Key header is supported. Same key returns the same job's current state. Response 202 json { "id": "dj_a1b2c3d4...", "status": "pending", "created_at": "2026-05-01T12:00:00Z" } Webhook events deep_search.completed : json { "id": "dj_a1b2c3d4...", "event_type": "deep_search.completed", "results": { "first_name": "Patrick", "last_name": "Collison", "full_name": "Patrick Collison", "current_job_title": "CEO", "headline": "CEO at Stripe", "location": "San Francisco, California, United States", "professional_url": "linkedin.com/in/patrickcollison", "company": { "name": "Stripe", "domain": "stripe.com", "professional_url": "linkedin.com/company/stripe", "industry": "Financial Services", "employee_count": 8000 }, "email": "patrick@stripe.com", "experience": {"company": "Stripe", "title": "CEO", "date_range": "2010–present"} , "score": 92 } , "total_returned": 1 } Company professional_url is optional and may be null . deep_search.failed : json { "id": "dj_a1b2c3d4...", "event_type": "deep_search.failed", "error": {"code": "error", "message": "Deep search failed; please retry."} } error.code error.message --- --- invalid_input Invalid deep-search query. error Deep search failed; please retry. A 0-result outcome ships as deep_search.completed with total_returned: 0 . Person and company professional_url values come back in bare-host form no scheme, no www. , no trailing slash . Company professional_url is optional and may be null . Billing Fixed charge per run. Customer pays the full mode price on a successful run regardless of total_returned . Pipeline errors and dispatch failures release the reservation no charge . Example bash curl -X POST https://api-public.clodo.ai/api/public/v1/deep-search/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "query": "Climate tech founders in Berlin", "mode": "standard", "webhook_url": "https://yourapp.com/webhooks/clodo" }' Rate limit POST /api/public/v1/deep-search/ : Limit Value --- --- Sustained rate 4 requests / minute Burst 2 requests Polling GET /api/public/v1/deep-search/{job_id}/ : Limit Value --- --- Sustained rate 3,600 requests / minute Burst 100 requests Exceeding any returns 429 . See also - Expert Discovery guide-experts - GTM Lead Generation guide-gtm - Recruiting & Candidate Sourcing guide-recruiting - Webhook Signing Secrets webhook-secrets - Webhook Signature Verification webhook-signing - Async Polling async-polling - Idempotency idempotency 

Deep Search

Find a professional email address from a profile URL or a name and company domain.

 Email Enrichment POST /api/public/v1/enrich/email/ — sync. 5 credits per call. Request Provide at least one of: - professional_url - first_name + last_name + company_name - first_name + last_name + domain Use professional_url , or first_name + last_name + domain , to resolve an email. The name + company_name shape is accepted for compatibility, but without a profile URL or domain it returns 404 not_found . Completed lookups that find no match still cost 5 credits. Field Type Notes --- --- --- professional_url string LinkedIn URL. first_name string Required when not using professional_url . last_name string Required when not using professional_url . company_name string Accepted with first_name + last_name , but cannot resolve an email without professional_url . Mutually exclusive with domain . domain string e.g. acme.com . Pair with first_name + last_name . Must contain a dot. Response json { "email": "patrick@stripe.com" } A miss returns 404 not_found see Status mapping below . Status mapping Outcome HTTP Charged --- --- --- Email found 200 OK 5 credits No match found 404 not_found 5 credits Invalid input 400 invalid_request 0 Upstream failure 502 upstream_error 0 Examples By LinkedIn URL: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/email/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{"professional_url": "https://www.linkedin.com/in/patrickcollison/"}' python import os import requests response = requests.post "https://api-public.clodo.ai/api/public/v1/enrich/email/", headers={"x-api-key": os.environ "CLODO_API_KEY" }, json={"professional_url": "https://www.linkedin.com/in/patrickcollison/"}, timeout=30, response.raise_for_status print response.json javascript const response = await fetch "https://api-public.clodo.ai/api/public/v1/enrich/email/", { method: "POST", headers: { "x-api-key": process.env.CLODO_API_KEY, "Content-Type": "application/json", }, body: JSON.stringify { professional_url: "https://www.linkedin.com/in/patrickcollison/", } , }, ; if !response.ok { throw new Error HTTP ${response.status}: ${await response.text } ; } console.log await response.json ; By profile URL with name + company context: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/email/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "professional_url": "https://www.linkedin.com/in/patrickcollison/", "first_name": "Patrick", "last_name": "Collison", "company_name": "Stripe" }' By name + domain: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/email/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "domain": "stripe.com" }' Rate limit Limit Value --- --- Sustained rate 30 requests / minute Burst 5 requests Exceeding any returns 429 . See also - Authentication authentication - Error Envelope errors - Credit Semantics credits 

Email Enrichment

Find a person’s phone number from their professional identity.

 Phone Enrichment POST /api/public/v1/enrich/phone/ — async returns 202 Accepted . 50 credits on hit, 0 on miss. Phone-number lookup. Typical completion time ranges from a few seconds to around a minute. Always returns 202 + webhook delivery for a uniform integration shape. Polling fallback at GET /api/public/v1/enrich/phone/{job_id}/ . Request Provide at least one of: - professional_url - email - first_name + last_name + company_name - first_name + last_name + domain Field Type Notes --- --- --- professional_url string LinkedIn URL. email string Person's professional email. first_name string Required when not using professional_url or email . last_name string Required when not using professional_url or email . company_name string Pair with first_name + last_name . Mutually exclusive with domain . domain string e.g. acme.com . Pair with first_name + last_name . Must contain a dot. webhook_url string Required. HTTPS URL. Must resolve to a public IP. Pre-condition: an active webhook signing secret. See Webhook Signing Secrets webhook-secrets . Idempotency-Key header is supported. Same key returns the same job's current state. Response 202 json { "id": "dj_a1b2c3d4...", "status": "pending", "created_at": "2026-05-01T12:00:00Z" } Webhook events phone_enrich.completed hit : json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.completed", "phone": "+14155551234" } phone_enrich.completed miss : json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.completed", "phone": null } phone_enrich.failed : json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.failed", "error": {"code": "error", "message": "Phone enrichment failed; please retry."} } error.code error.message --- --- invalid_input Invalid phone-enrichment request. error Phone enrichment failed; please retry. Branch on phone === null to detect miss, not on event_type . A miss is a clean completion. phone is E.164 normalized. Billing 50 credits charged on hit. 0 on miss. 0 on error or invalid input. Examples By LinkedIn URL: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/phone/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "professional_url": "https://www.linkedin.com/in/patrickcollison/", "webhook_url": "https://yourapp.com/webhooks/clodo" }' By email: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/phone/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "email": "patrick@stripe.com", "webhook_url": "https://yourapp.com/webhooks/clodo" }' By name + company: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/phone/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "company_name": "Stripe", "webhook_url": "https://yourapp.com/webhooks/clodo" }' By name + domain: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/phone/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "domain": "stripe.com", "webhook_url": "https://yourapp.com/webhooks/clodo" }' Rate limit POST /api/public/v1/enrich/phone/ : Limit Value --- --- Sustained rate 15 requests / minute Burst 3 requests Polling GET /api/public/v1/enrich/phone/{job_id}/ : Limit Value --- --- Sustained rate 3,600 requests / minute Burst 100 requests Exceeding any returns 429 . See also - Webhook Signing Secrets webhook-secrets - Webhook Signature Verification webhook-signing - Async Polling async-polling - Idempotency idempotency 

Phone Enrichment

Resolve a person’s identity to a professional profile URL.

 Professional URL Enrichment POST /api/public/v1/enrich/professional-url/ — sync. 3 credits per call. Resolve a person to their LinkedIn URL. Request Provide exactly one of: - email - first_name + last_name + company_name - first_name + last_name + domain These shapes are mutually exclusive. Mixing returns 400 invalid_request . Field Type Notes --- --- --- email string Standalone shape. first_name string Pair with last_name and company_name or domain . last_name string Pair with first_name and company_name or domain . company_name string Pair with first_name + last_name . Mutually exclusive with domain . domain string e.g. acme.com . Pair with first_name + last_name . Must contain a dot. Response json { "professional_url": "linkedin.com/in/patrickcollison" } A miss returns 404 not_found see Status mapping below . The URL comes back in bare-host form no scheme, no www. , no trailing slash . Status mapping Outcome HTTP Charged --- --- --- URL found 200 OK 3 credits No match found 404 not_found 3 credits Invalid input 400 invalid_request 0 Upstream failure 502 upstream_error 0 Examples By email: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/professional-url/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{"email": "patrick@stripe.com"}' By name + company: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/professional-url/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "company_name": "Stripe" }' By name + domain: bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/professional-url/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "first_name": "Patrick", "last_name": "Collison", "domain": "stripe.com" }' Rate limit Limit Value --- --- Sustained rate 6 requests / minute Burst 2 requests Exceeding any returns 429 . See also - Authentication authentication - Error Envelope errors - Credit Semantics credits 

Professional URL

Retrieve a person’s work history, education, and current role from their profile URL.

 Professional Profile Enrichment POST /api/public/v1/enrich/professional-profile/ — sync. 2 credits per call. Fetch a structured profile snapshot for a LinkedIn URL. Request Field Type Required Notes --- --- --- --- professional_url string yes LinkedIn URL. /in/<slug shape. Response json { "first_name": "Satya", "last_name": "Nadella", "full_name": "Satya Nadella", "headline": "Chairman and CEO at Microsoft", "location": { "city": "Redmond, Washington", "country": "United States", "country_code": "us" }, "current_position": { "title": "Chairman and CEO", "company_name": "Microsoft", "company_industry": "Computer Software" }, "experience": { "title": "Chairman and CEO", "company_name": "Microsoft", "start": "2014-02", "end": null } , "education": { "school_name": "University of Wisconsin-Milwaukee", "degree": "Master’s Degree", "field_of_study": "Computer Science", "start": null, "end": null } } Field shapes: - location : {city, country, country_code} . Any sub-field may be null . - current_position : {title, company_name, company_industry} , or null when the profile has no current position. - experience : {title, company_name, start, end} . start and end are YYYY-MM strings or null . - education : {school_name, degree, field_of_study, start, end} . Same date format. Any string field may be null when the source profile lacks it. Status mapping Outcome HTTP Charged --- --- --- Profile found 200 OK 2 credits No profile found 404 not_found 2 credits Invalid URL 400 invalid_request 0 Upstream failure 502 upstream_error 0 Example bash curl -X POST https://api-public.clodo.ai/api/public/v1/enrich/professional-profile/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{"professional_url": "https://www.linkedin.com/in/patrickcollison/"}' Rate limit Limit Value --- --- Sustained rate 18 requests / minute Burst 3 requests Exceeding any returns 429 . See also - Authentication authentication - Error Envelope errors - Credit Semantics credits 

Professional Profile

Draft personalized emails, review them, and send from your connected inbox.

 Outreach Emails Create AI-drafted outreach emails, review them, and send them from your own connected inbox — the full draft → review → approve → send loop over the API. 2 credits per drafted email , charged when the draft succeeds. Reviewing, editing, approving, cancelling, and reading are free. Emails send from a mailbox you've connected in the Clodo web app see Sending Accounts email-accounts , and every email you create here also appears on your Emails page in the app — same review queue, same thread history, same open and reply tracking. This is not a raw email-sending API: you supply a contact and instructions, Clodo drafts the copy. Every recipient is saved as a lead in your workspace. Lifecycle POST /emails/ - 202 status: drafting GET /emails/{id}/ - poll until status: pending_review PATCH /emails/{id}/ - optional draft edits POST /emails/{id}/approve/ - status: queued then sending - sent Status Meaning --- --- drafting The draft is being written typically 10–60s . pending_review Draft ready — waiting for your approval. Expires after 7 days see review_expires_at . queued Approved; waiting for its send slot. sending / sent In flight / delivered to your provider. failed Drafting or sending failed see error . Failed drafts are not charged. cancelled Cancelled by you, or expired unapproved. Pass "review": "none" at create to skip the review gate — the draft goes straight to queued with an automatically chosen send time. The default "required" is recommended: nothing sends without an explicit approve. Create — POST /emails/ Field Type Notes --- --- --- to.email string Required. Recipient address — becomes/updates a lead in your workspace. to.first_name string Required. to.last_name string Optional. to.company_name string Optional but recommended — improves the draft. to.professional_url string Optional LinkedIn URL. to.context string Optional free-text grounding, ≤1000 chars e.g. "raised Series B last month" . instructions string Required, ≤2000 chars. What the email should say/do. from_account string Optional ea_... id see Sending Accounts email-accounts . Defaults to your least-loaded connected account. reply_to string Optional Reply-To address. Must be one of your connected email accounts; anything else is a 400 . cc string Optional CC recipients, max 3. Visible to everyone on the thread. A reply from a CC'd address is not treated as the lead replying. bcc string Optional BCC recipients, max 3 e.g. a CRM logging address . Never visible to the other recipients. review string "required" default or "none" . bash curl -X POST https://api-public.clodo.ai/api/public/v1/emails/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "to": { "email": "jackie@acme.com", "first_name": "Jackie", "company_name": "Acme", "context": "spoke at SaaStr about PLG onboarding" }, "instructions": "Casual note asking if she is open to hearing about our founding PM role." }' Returns 202 : json { "id": "em_1f0c9a...", "status": "drafting", "created_at": "2026-07-13T00:00:00Z" } 409 conflict means you have no usable sending account — connect one in the Clodo web app first. Read — GET /emails/{id}/ Poll this while drafting ; afterwards it's the live state of the email opens, replies, schedule . json { "id": "em_1f0c9a...", "status": "pending_review", "to": { "email": "jackie@acme.com", "name": "Jackie" }, "from_email": "you@yourcompany.com", "subject": "Founding PM at Acme?", "body_plain": "Hi Jackie, ...", "body_html": "<p Hi Jackie, ...</p ", "instructions": "Casual note asking...", "created_at": "2026-07-13T00:00:00Z", "scheduled_for": null, "sent_at": null, "opened_at": null, "open_count": 0, "replied_at": null, "review_expires_at": "2026-07-20T00:00:10Z", "error": null } List — GET /emails/ Paginated, newest first. Filters: ?status=<any wire status , ?since=<ISO 8601 , ?page= , ?page_size= default 50, max 200 . Returns {count, page, page_size, total_pages, results} . Only emails created through the API appear here. bash curl "https://api-public.clodo.ai/api/public/v1/emails/?status=replied&since=2026-07-01T00:00:00Z" \ -H "x-api-key: ck_live_..." Edit — PATCH /emails/{id}/ Allowed while pending_review or queued approved but not yet sent — including scheduled follow-up steps of a running sequence . Send subject and/or one of body_plain / body_html not both . HTML is sanitized server-side. Editing while pending_review extends the 7-day review window. Once a message is sending or sent the edit returns 409 . bash curl -X PATCH https://api-public.clodo.ai/api/public/v1/emails/em_1f0c9a.../ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{"subject": "Quick question about Acme"}' Approve — POST /emails/{id}/approve/ Only while pending_review . Returns the updated email. Body Behavior --- --- {"schedule": "best_time"} Default. Clodo picks the optimal send time for the recipient. {"schedule": "now"} Next available send slot. {"schedule": "at", "send_at": "2026-07-14T16:00:00Z"} Fixed UTC time. Cancel — POST /emails/{id}/cancel/ Allowed while drafting , pending_review , or queued . Cancelling mid-draft costs nothing; cancelling an already-drafted email does not refund the 2 credits the drafting work happened . Status mapping Outcome HTTP Charged --- --- --- Draft created 202 2 credits when the draft succeeds Draft failed — status: failed on poll 0 Invalid input 400 invalid_request 0 No usable sending account 409 conflict 0 Edit/approve/cancel in wrong state 409 conflict 0 Unknown em_ id 404 not_found 0 Notes - Webhooks are optional for these endpoints — polling is first-class. If you never register a webhook secret, everything above still works. - Daily send limits on your connected accounts always apply; approving more than the day's limit queues the remainder for the following days. - Sends stop automatically when the recipient replies visible via replied_at . See also - Sending Accounts email-accounts - Credit Semantics credits - Async Polling async-polling 

Outreach Emails

Create, review, and run personalized email sequences with automatic follow-ups.

 Sequences Enroll up to 100 contacts in a multi-step drafted sequence — every step is drafted up front, you review the whole thing, and one approve sets it running. 2 credits per drafted email e.g. 5 contacts × 3 steps = 30 credits , reserved when you create the sequence and settled when drafting finishes; failed drafts are released. Sequences require a connected Google or Microsoft sending account reply detection drives stop-on-reply; SMTP can't support it — you'll get a 409 . How follow-ups behave - Step timing is automatic: follow-ups go out 3 / 7 / 14 days after the previous step actually sends tunable per-account in the web app , at a sensible hour, skipping weekends. - A step only sends after the prior step has sent. - A reply or bounce stops everything : remaining steps for that contact are cancelled automatically. - Each enrolled contact becomes a lead in your workspace, and the whole sequence is visible on your Emails page in the app. Lifecycle POST /sequences/ - 202 status: drafting GET /sequences/{id}/ - poll until status: pending_review PATCH /emails/{em_id}/ - optional per-step edits steps are regular em_ emails POST /sequences/{id}/approve/ - status: active steps queued on cadence POST /sequences/{id}/contacts/- enroll more contacts, any time after drafting settles GET /sequences/{id}/replies/ - inbound replies as they arrive Sequence status Meaning --- --- drafting Steps are being written ~30–60s per email . pending_review All drafts ready — waiting for approve. active Approved and running. completed / cancelled / failed Terminal. Per-enrollment statuses carry the detail replied , bounced , ... . Create — POST /sequences/ Field Type Notes --- --- --- contacts array, 1–100 Same contact shape as Outreach Emails endpoint-outreach-emails email + first_name required; company_name , professional_url , context optional . instructions string, ≤2000 What the sequence should pitch/do. name string, ≤200 Optional label for your own bookkeeping e.g. "LegalHR1" . Echoed on every sequence payload; not shown to recipients. steps int, 1–4 Total steps including the intro. Default 3. from_account string Optional ea_... sender pin. reply_to string Optional Reply-To address for every step. Must be one of your connected email accounts any of them — not just the sender ; anything else is a 400 . Replies routed there are still detected: the sequence stops on reply as usual. cc string Optional CC recipients, max 3, applied to every step. Visible to everyone on the thread. A reply from a CC'd address does not stop the sequence — only the lead's reply does. bcc string Optional BCC recipients, max 3, applied to every step e.g. a CRM logging address . Never visible to the other recipients. review string "required" default or "none" auto-approve after drafting . bash curl -X POST https://api-public.clodo.ai/api/public/v1/sequences/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{ "contacts": {"email": "x@acme.com", "first_name": "Xola", "company_name": "Acme"}, {"email": "y@globex.com", "first_name": "Yuri", "company_name": "Globex"}, {"email": "z@initech.io", "first_name": "Zadie", "company_name": "Initech"} , "instructions": "3-step sequence pitching our outbound API. Casual tone, developer audience.", "name": "DevRel-Q3-1", "steps": 3 }' Returns 202 : json { "id": "sq_9c41d2...", "name": "DevRel-Q3-1", "status": "drafting", "created_at": "2026-07-13T00:00:00Z", "contacts_accepted": 3, "contacts_rejected": , "total_emails": 9, "drafted": 0 } Partial accept: ineligible contacts are skipped, not fatal — contacts_rejected lists each with a reason already_in_automation — the person is in another active sequence; not_enriched ; recently_contacted ; invalid_contact . Only if no contact is eligible does the call fail. Add contacts — POST /sequences/{id}/contacts/ Enroll 1–100 additional contacts into an existing sequence — including after it's active. New enrollments inherit the sequence's original instructions and step plan and start at step 0 on their own cadence clock; existing enrollments are untouched. 2 credits per newly drafted email , reserved on the call and settled when the new drafts finish. Field Type Notes --- --- --- contacts array, 1–100 Same contact shape as create. from_account string Optional ea_... sender pin for the new enrollments. New enrollments also inherit the sequence's reply_to , cc , and bcc if set at create time — appended contacts' threads behave the same as everyone else's. review string "required" default or "none" . "none" auto-approves every pending enrollment in the sequence once drafting finishes. bash curl -X POST https://api-public.clodo.ai/api/public/v1/sequences/sq_9c41d2.../contacts/ \ -H "x-api-key: ck_live_..." \ -H "Content-Type: application/json" \ -d '{"contacts": {"email": "w@umbrella.co", "first_name": "Wren", "company_name": "Umbrella"} }' Returns 202 : json { "id": "sq_9c41d2...", "name": "DevRel-Q3-1", "status": "drafting", "contacts_accepted": 1, "contacts_rejected": , "emails_added": 3 } Behavior notes: - The sequence's top-level status returns to drafting while the new enrollments draft, then settles back per-enrollment statuses stay accurate throughout — poll the detail . - With review: "required" , call approve/ again after drafting — it activates only the newly drafted enrollments already-active ones are skipped . Per-step copy edits via PATCH /emails/{em_id}/ work on the new drafts before approval, same as at create. - 409 while a drafting pass is running retry shortly , and for cancelled / failed sequences create a new one instead . - Contacts already enrolled in this sequence are rejected with reason already_enrolled ; the other rejection reasons match create. - A sequence holds at most 500 enrollments. Read — GET /sequences/{id}/ json { "id": "sq_9c41d2...", "name": "DevRel-Q3-1", "status": "pending_review", "instructions": "3-step sequence pitching...", "total_emails": 9, "drafted": 9, "failed": 0, "enrollments": { "contact": {"email": "x@acme.com", "name": "Xola"}, "status": "pending_review", "emails": {"id": "em_...", "step": 0, "status": "pending_review", "subject": "...", "body_plain": "...", ...}, {"id": "em_...", "step": 1, ...}, {"id": "em_...", "step": 2, ...} } } Each step is a regular outreach email — the em_ ids work against GET /emails/{id}/ and PATCH /emails/{id}/ for per-step edits. Edits work before approving AND after — any step whose status is pending_review or queued accepts a PATCH , so you can rewrite upcoming follow-ups on a live sequence. Every other status drafting , sending , sent , failed , cancelled returns 409 . GET /sequences/ lists your sequences paginated, no enrollments — fetch the detail . Approve — POST /sequences/{id}/approve/ Approves every drafted enrollment: step 0 queues at its computed best time, later steps follow the cadence after each prior send. 409 while still drafting. Cancel — POST /sequences/{id}/cancel/ Cancels all non-terminal enrollments and their unsent steps. Already-sent steps are unaffected. No refunds for drafted emails the drafting work happened . Replies — GET /sequences/{id}/replies/ json { "replies": { "email_id": "em_...", "from_email": "x@acme.com", "from_name": "Xola", "subject": "Re: your outbound API", "body_plain": "Interested — got time Thursday?", "classification": "reply", "received_at": "2026-07-16T14:03:00Z" } } classification is reply , bounce , or auto_reply . The corresponding enrollment will already show status replied / bounced with its remaining steps cancelled. Status mapping Outcome HTTP Charged --- --- --- Sequence created 202 2 credits × drafts that succeed All contacts rejected 400 invalid_request 0 No Google/Microsoft account 409 conflict 0 Approve while drafting 409 conflict 0 Unknown sq_ id 404 not_found 0 See also - Outreach Emails endpoint-outreach-emails — the one-off flow and per-step edit semantics - Sending Accounts email-accounts - Credit Semantics credits 

Sequences

Check your connected mailboxes before creating outreach emails or sequences.

 Sending Accounts GET /api/public/v1/email/accounts/ — sync. Free. Outreach emails send from your own connected mailbox Google, Microsoft, or SMTP — your address, your sending reputation. Accounts are connected and managed in the Clodo web app Settings → Email Accounts ; the API surface is read-only. Use this endpoint to check that a usable inbox exists before creating emails. Response json { "accounts": { "id": "ea_65006cf5b9b8c2af", "email": "you@yourcompany.com", "type": "google", "connected": true, "sending_enabled": true, "daily_limit": 35, "sent_today": 12, "health": "healthy" } } Field Notes --- --- id Stable ea_... handle. Pass as from_account when creating an email to pin the sender. type google , microsoft , or smtp . connected false means the account needs attention in the web app e.g. re-auth . daily_limit Effective sends allowed today accounts warm up gradually and throttle on poor engagement . health healthy , or a state describing what needs fixing e.g. reauth_required . If the list is empty or nothing is connected + sending_enabled , POST /emails/ returns 409 conflict — connect an inbox in the web app first. See also - Outreach Emails endpoint-outreach-emails 

Sending Accounts

Create and manage the signing secrets used to authenticate webhook deliveries.

 Webhook Signing Secrets Async endpoints People Search, Deep Search, Phone Enrichment POST results to a webhook URL you supply per request. Each delivery is signed with HMAC-SHA256 using your webhook signing secret. An active webhook signing secret is required to call any async endpoint. Calls without one fail with 409 conflict . Minting 1. Open the API Keys tab. 2. Find the Webhook Signing Secret section. 3. Click Mint webhook secret . 4. Copy the raw secret. It is shown once. The Console displays the secret's prefix and version after creation. Rotating Click Rotate to mint a new secret and invalidate the old one immediately. API key vs webhook signing secret Thing Used for Wire format --- --- --- API key ck_live_... Authenticating your requests to us x-api-key header on outbound requests Webhook signing secret Verifying webhook deliveries from us to you Clodo-Signature header on inbound webhook POSTs Scope One active webhook signing secret per account. All webhook deliveries to the account use it. Headers we send on every webhook Header Value --- --- User-Agent Clodo-Webhook/1.0 Content-Type application/json Clodo-Signature t=<unix ,v1=<hmac_sha256_hex HMAC over <t .<body Clodo-Webhook-Event Event type e.g. search.completed . Lets you route on a header without parsing the body. Clodo-Webhook-Id Numeric delivery ID. Stable across retry attempts of the same delivery. Wire format The body bytes are canonicalized: keys sorted alphabetically, no whitespace. Sign and verify the bytes you receive on the wire, not a re-serialized JSON dict key order would diverge and HMAC would fail . What to read next - Webhook Signature Verification webhook-signing for the HMAC algorithm and verifier code. - Webhook Events webhook-events for event types and payload shapes. - Webhook Retry Policy webhook-retries for retry schedule. 

Signing Secrets

Verify that a webhook came from Clodo before processing its payload.

 Webhook Signature Verification Every webhook delivery includes a Clodo-Signature header. Verify it against the raw request body using your webhook signing secret. Header format Clodo-Signature: t=<unix_timestamp ,v1=<hmac_sha256_hex - t — unix seconds when we signed the request. - v1 — lowercase hex HMAC-SHA256 over <t .<body_bytes , keyed with your secret. Algorithm 1. Read the raw bytes of the request body. Do not parse and re-serialize — key order would change and the HMAC would not match. 2. Parse t and v1 from the Clodo-Signature header. 3. Reject if abs now - t 300 5-minute replay window . 4. Compute expected = HMAC_SHA256 secret_utf8, f"{t}.".encode "ascii" + body_bytes .hexdigest . 5. Constant-time compare expected with v1 . Python verifier python import hashlib import hmac import time TOLERANCE_SECONDS = 5 60 def verify secret: str, body: bytes, signature_header: str - bool: parts = dict p.split "=", 1 for p in signature_header.split "," if "t" not in parts or "v1" not in parts: return False try: t = int parts "t" except ValueError: return False if abs int time.time - t TOLERANCE_SECONDS: return False signed = f"{t}.".encode "ascii" + body expected = hmac.new secret.encode "utf-8" , signed, hashlib.sha256 .hexdigest return hmac.compare_digest expected, parts "v1" Flask example: python @app.post "/webhooks/clodo" def receive : body = request.get_data raw bytes if not verify SECRET, body, request.headers.get "Clodo-Signature", "" : return "", 401 payload = json.loads body ... handle payload ... return "", 200 Node verifier javascript import crypto from "node:crypto"; const TOLERANCE_SECONDS = 5 60; export function verify secret, body, signatureHeader { const parts = Object.fromEntries signatureHeader.split "," .map p = p.split "=", 2 , ; if !parts.t !parts.v1 return false; const t = parseInt parts.t, 10 ; if !Number.isFinite t return false; if Math.abs Math.floor Date.now / 1000 - t TOLERANCE_SECONDS return false; const signed = Buffer.concat Buffer.from ${t}. , "ascii" , body ; const expected = crypto .createHmac "sha256", secret .update signed .digest "hex" ; const a = Buffer.from expected, "hex" ; const b = Buffer.from parts.v1, "hex" ; return a.length === b.length && crypto.timingSafeEqual a, b ; } Express example use express.raw so req.body stays as bytes : javascript app.post "/webhooks/clodo", express.raw { type: "application/json" } , req, res = { const ok = verify SECRET, req.body, req.get "Clodo-Signature" "" ; if !ok return res.status 401 .end ; const payload = JSON.parse req.body.toString "utf8" ; // ... handle payload ... res.status 200 .end ; }, ; Common mistakes - Parsing the JSON before verifying. Most frameworks consume the body as a stream; once consumed it is gone. Use raw-body middleware express.raw , request.get_data , etc. and verify against the bytes. - Comparing strings with == . Use hmac.compare_digest Python or crypto.timingSafeEqual Node . - Trusting t without bounds-checking. The 5-minute window is the replay defense. See also - Webhook Signing Secrets webhook-secrets for minting and rotating the secret. - Webhook Events webhook-events for the catalog of events and their payload shapes. - Webhook Retry Policy webhook-retries for retry schedule. 

Signature Verification

Reference payloads for search and enrichment completion and failure events.

 Webhook Events Catalog of every event we deliver. All bodies are alphabetically key-sorted + minified on the wire — see Webhook Signing Secrets webhook-secrets . search.completed Fired when a People Search run finishes successfully including 0-result runs . json { "id": "dj_a1b2c3d4...", "event_type": "search.completed", "results": { "first_name": "Mark", "last_name": "Wang", "full_name": "Mark Wang", "current_job_title": "Software Engineer", "headline": "Software Engineer", "location": "San Francisco, California, United States", "professional_url": "linkedin.com/in/mark-wang-8a8a2213", "company": { "name": "OpenAI", "domain": "openai.com", "professional_url": "linkedin.com/company/openai", "industry": "Research Services", "employee_count": 8626 } } , "total_returned": 1 } Company professional_url is optional and may be null . A 0-result outcome ships with total_returned: 0 and results: . search.failed Fired when a People Search run fails before producing results. json { "id": "dj_a1b2c3d4...", "event_type": "search.failed", "error": {"code": "error", "message": "People search failed; please retry."} } error.code error.message --- --- invalid_input Invalid search query. error People search failed; please retry. deep_search.completed Fired when a Deep Search run finishes successfully including 0-result runs . json { "id": "dj_a1b2c3d4...", "event_type": "deep_search.completed", "results": { "first_name": "Patrick", "last_name": "Collison", "full_name": "Patrick Collison", "current_job_title": "CEO", "headline": "CEO at Stripe", "location": "San Francisco, California, United States", "professional_url": "linkedin.com/in/patrickcollison", "company": { "name": "Stripe", "domain": "stripe.com", "professional_url": "linkedin.com/company/stripe", "industry": "Financial Services", "employee_count": 8000 }, "email": "patrick@stripe.com", "experience": {"company": "Stripe", "title": "CEO", "date_range": "2010–present"} , "score": 92 } , "total_returned": 1 } email , headline , experience , and company.professional_url may be empty "" , , or null on a per-lead basis. Person and company professional_url values use bare-host form. deep_search.failed json { "id": "dj_a1b2c3d4...", "event_type": "deep_search.failed", "error": {"code": "error", "message": "Deep search failed; please retry."} } error.code error.message --- --- invalid_input Invalid deep-search query. error Deep search failed; please retry. agentic_search.completed Fired when an Agentic Search finishes, including zero-result and partial deliveries. The payload contains id , event_type: "agentic_search.completed" , results , total_returned , partial , and completion_reason . Each person may include professional details, company, experience, fit_summary , proof_points , evidence , and tier . See Agentic Search endpoint-agentic-search for the complete example and field descriptions. When partial is true , inspect completion_reason : budget_exhausted , deadline_exceeded , or user_cancelled . A partial delivery can still be billable within the requested credit cap. agentic_search.failed json { "id": "dj_a1b2c3d4e5f6478890abcdef123456789", "event_type": "agentic_search.failed", "error": {"code": "error", "message": "Internal error — you were not charged."} } error.code Meaning --- --- error Internal failure; no charge is retained. stale_running_1h Job timed out and was force-failed; the credit hold was refunded. admission_expired The job's billing authorization is no longer valid. Check billing and submit a new request. settlement_failed Billing could not settle the request. No result is delivered and no charge is retained. See Async Polling async-polling if webhook delivery is unavailable. phone_enrich.completed Fired for both hits and clean misses. Branch on phone === null to detect a miss. Hit: json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.completed", "phone": "+14155551234" } Miss: json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.completed", "phone": null } phone is E.164 normalized. phone_enrich.failed Fired only on transport / pipeline errors. json { "id": "dj_a1b2c3d4...", "event_type": "phone_enrich.failed", "error": {"code": "error", "message": "Phone enrichment failed; please retry."} } error.code error.message --- --- invalid_input Invalid phone-enrichment request. error Phone enrichment failed; please retry. See also - Webhook Signature Verification webhook-signing for HMAC verifier code. - Webhook Retry Policy webhook-retries for retry schedule and what we treat as transient vs permanent. - Async Polling async-polling for the polling-fallback envelope. 

Webhook Events

Understand delivery attempts, acknowledgements, and duplicate events.

 Webhook Retry Policy Schedule Three attempts max. Schedule: Attempt When --- --- 1 At result-ready time 2 5 seconds after attempt 1 fails 3 5 minutes after attempt 2 fails After attempt 3 fails, the delivery is marked failed and not retried. A 1-hour hard ceiling from the first attempt also caps total time. If a delivery is still pending past 1 hour after first attempt, it is force-failed. Success Any 2xx response from your endpoint marks the delivery delivered . We do not retry on 2xx. What we retry Transient failures get re-enqueued per the schedule above: - 5xx HTTP responses - Connection errors DNS resolution, refused, reset, peer reset - Read or connect timeouts 5s connect / 20s read per attempt - Other transport-level errors What we do NOT retry Permanent failures move straight to failed : - 4xx HTTP responses treated as "your code rejected this delivery; retrying won't help" - TLS / certificate errors expired, hostname mismatch, custom CA - SSRF rejections at delivery time URL resolved to a private / loopback / IMDS address - Your account has no active webhook signing secret Headers each retry sends Every attempt for a given delivery sends the same Clodo-Webhook-Id header. Use that value on your end to dedup if you process and a 2xx response is lost in flight. The Clodo-Signature is recomputed per attempt with a fresh t= . Within the 5-minute replay window the signature stays valid against the same body bytes. Reconciling lost deliveries If your receiver is down past the retry window, the delivery is failed but the underlying job result is still durable. Use Async Polling async-polling to fetch the result via GET /v1/<endpoint /{job_id}/ . The polling response carries the same body as the webhook would have delivered without the id / event_type wrapper . See also - Webhook Signing Secrets webhook-secrets - Webhook Signature Verification webhook-signing - Webhook Events webhook-events - Async Polling async-polling 

Retry Policy

Build B2B lead generation, candidate sourcing, and expert discovery workflows with Clodo. Compare search endpoints and connect research, enrichment, and outreach.

 People Search API Use Cases Build workflows for B2B lead generation, candidate sourcing, and expert discovery with the Clodo API. Start by describing the people you need, choose the search endpoint that fits the task, enrich selected profiles, and prepare outreach from a connected mailbox. Describe the exact person you have in mind Get specific about the combination that matters: someone at a company using a particular technology, who has solved a particular problem, worked in a particular environment, or published or presented relevant work. Include examples, exclusions, priorities, and dates. An ICP or candidate persona can describe a story about the person's experience, not just a list of profile attributes. For example: “Find fintech data leaders at companies using Snowflake who previously built banking infrastructure. Prioritize people who have given a talk about a migration since January 2025, and include supporting sources.” Agentic Search endpoint-agentic-search is the starting point for this level of detail. Review the returned evidence against your requirements and keep unverified criteria as open questions. Choose a workflow Goal Guide What you will build --- --- --- Find prospects for your product GTM Lead Generation guide-gtm Turn an ideal customer profile into a prospect list, resolve email addresses, and draft relevant sales outreach. Source people for an open role Recruiting & Candidate Sourcing guide-recruiting Translate a job description into search criteria, research candidates, and prepare recruiter outreach. Find people with first-hand expertise Expert Discovery guide-experts Identify practitioners for research interviews, advisory projects, and industry discussions, then review their experience. Which people search endpoint should I use? Endpoint Best starting point How results help --- --- --- Agentic Search endpoint-agentic-search A detailed combination of company tech stack, professional experience, accomplishments, public work, priorities, and exclusions Researches matches and returns fit summaries, proof points, and source evidence. Set a maximum credit budget. Account enrollment is required. People Search endpoint-people-search A professional audience described by role, seniority, company, industry, or geography Returns matching professional profiles. Continue a completed search to retrieve more non-overlapping results. Deep Search endpoint-deep-search A focused search that needs a researched shortlist Returns professional context, work experience, and fit scores, with standard and extended modes priced per successful run. Use the endpoint that fits the information you need. You do not need to call all three for every workflow. If a broad list answers your question, start with People Search. If the distinction between a good and a poor match depends on specific experience, describe it explicitly in Agentic Search or Deep Search. Before you start 1. Create your account and get an API key get-api-key . Run API calls on your server and send the key in the x-api-key header. 2. Create a webhook signing secret webhook-secrets and provide a public HTTPS receiver for search jobs. Verify webhook signatures webhook-signing before processing results. 3. Store each accepted job ID. Use Async Polling async-polling if you need to retrieve its status or results directly. 4. Check Credits & Pricing credits before choosing result counts, search modes, or an Agentic Search cap. 5. If you plan to draft outreach, connect a mailbox and check Sending Accounts email-accounts . Search alone does not require a sending account. Connect the steps in your application Search → review matches → enrich selected people → draft outreach → approve sending. Save the search criteria alongside each job ID so you can explain why a person was included. On completion, retain the professional URL, company, and relevant search context. Use those fields to deduplicate your own records and decide who needs further research. Professional Profile Enrichment endpoint-professional-profile retrieves structured work history and education. Email Enrichment endpoint-email resolves an email from a professional URL or a name and company domain. Enrich only the people you intend to review or contact; search results do not guarantee a usable email address. Outreach Emails endpoint-outreach-emails turns recipient details and instructions into a draft. Set review: "required" , inspect the draft, and approve it when ready. For follow-up workflows, see Sequences endpoint-sequences . Questions about building people discovery workflows Can I use one API for lead generation and recruiting? Yes. The search endpoints accept descriptions of people rather than a fixed sales-only or recruiting-only request format. Change the query and your downstream review criteria to suit the workflow. The guides show different examples for each audience. Does finding someone automatically contact them? No. Search and enrichment do not send emails. Outreach requires a separate request and a connected sending account. With review: "required" , a draft waits for an explicit approval before entering the send queue. Can I connect results to my CRM, ATS, or research database? Your application can map returned JSON into its own records and write to those systems through their supported APIs. These guides describe that application step; the search endpoints do not automatically create CRM records, ATS candidates, or research projects. 

People Search API Use Cases

Learn how to find B2B sales leads with an API: turn your ICP into a prospect list, qualify decision-makers, enrich email addresses, and prepare personalized outreach.

 GTM Lead Generation with the Clodo API Turn an ideal customer profile into a list of people to contact. This guide connects natural-language people search, lead qualification, email enrichment, and personalized outreach for sales prospecting and go-to-market workflows. Your ICP can be as specific as the problem your product solves. Combine the person's responsibilities, their company's technology stack, past experience, and relevant public work. For example, look for an engineering leader at a fintech using Snowflake who previously built data infrastructure at a bank and has spoken publicly about data-platform migrations. 1. Translate your ICP into a people search Describe the people you would want to meet if you could write the exact criteria yourself. Start with who owns the problem, then explain the combination of company context, technology, experience, and public evidence that would make them a strong fit. Include must-haves, preferences, exclusions, and a time window when recency matters. You do not need to reduce a nuanced persona to a job title and location. “People who have done X, at companies using Y, with evidence of Z” is a useful way to write an Agentic Search endpoint-agentic-search query. Its query field supports up to 2,000 characters, so use that space to explain what a good match looks like. Starting point More useful search --- --- Manufacturing leads Operations leaders at US manufacturers who have taken warehouse automation from pilot to production; prioritize people who have described the implementation in a public case study Prospects for data infrastructure Data-platform leaders at fintech companies using Snowflake, with prior experience building data infrastructure at a bank Prospects for developer tooling Engineering leaders at B2B software companies using Kubernetes who have given a conference talk about platform engineering; exclude agencies and consultants Security decision-makers Security leaders at UK financial-services companies who previously led a cloud-security program and have published an article or spoken about its implementation Make the evidence requirement explicit: identify the technology, experience, publication, or event that matters, and ask for sources supporting the match. A query expresses what you want found; review the returned evidence to establish which criteria were actually verified. Missing evidence is an unanswered qualification question. Publicly accessible posts can also provide context: ask for people who have written about a particular implementation problem, and include a known post URL when you have one. A person's own public writing and the list of people who liked a post are different sources; the latter requires social-engagement access beyond the current public API. Example: combine tech-stack signals, career history, and public work For an enrolled account, send this body to POST /api/public/v1/agentic-search/ after completing the setup checklist guide-use-cases . Replace the example webhook URL with your public HTTPS receiver. json { "query": "Find data-platform leaders at US or UK fintech companies using Snowflake. Prioritize people who previously built data infrastructure at a bank and have written an engineering article or given a conference talk about a data-platform migration since January 2025. Exclude consultants and agencies. Company use of Snowflake is required; prior banking experience and public migration work are preferences. Include sources supporting the technology and experience matches, and distinguish missing evidence from confirmed facts.", "target_results": 25, "max_credits": 500, "webhook_url": "https://your-app.example/webhooks/clodo" } Wait for agentic_search.completed , or poll GET /api/public/v1/agentic-search/{job_id}/ . Review fit_summary , proof_points , and evidence , together with partial and completion_reason . Carry the supported context into qualification and outreach. You can proceed directly to step 3 with these results. 2. Build the initial prospect list For a broader first pass, use People Search endpoint-people-search . This is an alternative starting point to the detailed Agentic Search above, useful when a role and company description is enough to build your initial pool. After completing the setup checklist guide-use-cases , send this JSON body to POST /api/public/v1/search/ with your x-api-key header. Replace the example receiver with your public HTTPS webhook URL. json { "query": "Plant managers and operations directors at industrial manufacturing companies in the United States with 200 to 2000 employees.", "max_results": 50, "webhook_url": "https://your-app.example/webhooks/clodo" } Store the id from the 202 response. Wait for search.completed , or poll GET /api/public/v1/search/{job_id}/ . A completed webhook contains results ; a completed polling response places them in result.results . Inspect failures and terminal status before treating a job as ready. Need a researched shortlist? Deep Search endpoint-deep-search adds professional context and fit scores. For the detailed combination of technology, experience, and public work shown above, use Agentic Search endpoint-agentic-search and assess the evidence for each criterion. 3. Qualify leads before enriching contact details Keep each prospect's name, professional URL, current role, and company in your application. Compare them with your ICP and remove people outside your target audience. Your own CRM can hold the review status and campaign membership; Clodo search does not write those records automatically. Prioritize people whose responsibilities relate to the problem your product solves. A matching title is a starting point for qualification, not proof of budget or purchase intent. When more background would change your decision, call Professional Profile Enrichment endpoint-professional-profile to inspect work history. 4. Find an email for a selected prospect Send the person's returned professional_url to POST /api/public/v1/enrich/email/ . This is a synchronous request. The value below is a placeholder; use the actual profile URL you selected. json { "professional_url": "https://www.linkedin.com/in/selected-prospect" } A 200 returns email . A 404 means no email was resolved; leave the contact unresolved instead of inventing an address. Completed email lookups cost 5 credits even when there is no match, so enrich after qualification. See Email Enrichment endpoint-email for accepted input combinations and other errors. 5. Draft a relevant first touch Connect a mailbox and verify it appears in Sending Accounts email-accounts . Send a separate request to POST /api/public/v1/emails/ , using the resolved email and accurate recipient details. The example below shows the structure, not a real recipient. json { "to": { "email": "alex@manufacturer.example", "first_name": "Alex", "company_name": "Example Manufacturing", "context": "Operations director at an industrial manufacturer." }, "instructions": "Write a short introduction to our production-planning software. Connect it to the recipient's operations role without assuming they have a particular problem. Ask whether reducing manual scheduling work is a current priority. Do not claim we have met or invent company news.", "review": "required" } Poll GET /api/public/v1/emails/{id}/ until the email reaches pending_review or another terminal outcome. Review the subject and body, edit if needed, then call POST /api/public/v1/emails/{id}/approve/ to queue it for sending. Approval is a separate action; creating this draft does not send it. See Outreach Emails endpoint-outreach-emails for the full lifecycle and Sequences endpoint-sequences for follow-ups. 6. Expand the campaign without repeating the same page To continue a completed People Search, send its job ID as search_id with max_results and webhook_url . Omit query for the continuation. Clodo returns the next non-overlapping page for that search. Still deduplicate records in your own application when combining different searches or campaigns. Use Idempotency idempotency for retrying search requests and respect Rate Limits rate-limits . Keep search, enrichment, and drafting charges separate in your budget; a list of 50 prospects does not require 50 enrichment calls or 50 drafts. Lead generation questions How do I find B2B sales leads with an API? Describe the decision-maker and target company in People Search, review returned profiles against your ICP, then enrich contact details for the prospects you select. This separates audience discovery from the decision to contact someone. Can I search for decision-makers at target accounts? Include the company or account criteria and relevant functions in your query. Review returned company and role fields before adding someone to an account-based sales campaign. For complex qualification, use Agentic Search or Deep Search. How specific can my ideal customer profile be? Combine multiple criteria in plain language: company technology, current responsibilities, previous employers, implementation experience, published work, conference participation, geography, and recency. Use Agentic Search for a detailed persona and state which conditions are required versus preferred. Specificity guides the research; it does not guarantee that every condition has a publicly verifiable match. Does this workflow identify buying intent? A role or company match does not establish buying intent. If a particular public activity matters, describe it in an Agentic Search query and assess any supporting evidence. Do not treat an unsupported assumption as a verified signal. Related workflows - Recruiting & Candidate Sourcing guide-recruiting : build talent pipelines rather than prospect lists. - Expert Discovery guide-experts : find practitioners for interviews and advisory projects. - Credits & Pricing credits : plan the cost of each step. 

GTM Lead Generation

Source candidates from a job description with the Clodo API. Research professional experience, build recruiting shortlists, enrich profiles, and draft recruiter outreach.

 Recruiting and Candidate Sourcing with the Clodo API Build a candidate sourcing workflow that starts with a role description and ends with a reviewed shortlist and recruiter outreach. Use Clodo to find people with relevant professional experience, inspect their backgrounds, and explain why they may fit the role. This example searches for engineering leaders who have brought robotics systems into production. The same flow works for specialized individual contributors, industry operators, and leadership searches when you change the job-related criteria. 1. Turn a job description into a searchable persona Separate requirements from preferences. Tell the search what the person needs to have done, the environment they worked in, and which experience is optional. Avoid relying on a title alone: two engineering managers can have very different technical and operating backgrounds. Get specific enough to describe the person your team would be excited to meet: “An engineer who has operated Kubernetes in a regulated financial-services environment, previously worked at an early-stage company, and has presented a production incident case study.” Combine technical experience, career history, and public work; explain which are requirements and which would strengthen a match. Ask for evidence of the actual work, since employment at a company using a technology does not establish personal experience with it. Hiring requirement Search detail --- --- Relevant technical experience Deployed autonomous robotics systems in a production environment Leadership background Led an engineering team through a product launch Location Based in the United States Preference Prior warehouse or industrial robotics experience Keep the query focused on role-relevant evidence. Search results support sourcing and professional review; they do not establish availability, interest, or willingness to relocate. 2. Search for candidates with specific experience Agentic Search endpoint-agentic-search is the canonical endpoint for a detailed persona. It returns fit summaries, proof points, and evidence that help a recruiter understand the match. It requires account enrollment; see its availability section before starting. Complete the API and webhook setup guide-use-cases , then send this body to POST /api/public/v1/agentic-search/ with x-api-key . Replace the example webhook URL with your receiver. json { "query": "Find US-based engineering leaders who have taken autonomous robotics systems from prototype to production and led an engineering team through a product launch. Prioritize warehouse or industrial robotics experience. Exclude recruiting firms and consultants. Explain the relevant experience with supporting evidence.", "target_results": 25, "max_credits": 500, "webhook_url": "https://your-app.example/webhooks/clodo" } Save the returned job ID. Wait for agentic_search.completed , or poll GET /api/public/v1/agentic-search/{job_id}/ . The requested result count is a target, not a guarantee. Check partial and completion_reason before treating the search as complete in your recruiting workflow. For a broader candidate pool defined by title, industry, and location, use People Search endpoint-people-search . For a researched shortlist with fit scores and fixed pricing per successful run, use Deep Search endpoint-deep-search . 3. Build an evidence-based candidate review For each result, retain the professional URL, current role, company, fit_summary , proof_points , and evidence . In your own recruiting application, show the original requirement next to the supporting information so reviewers can distinguish a demonstrated match from missing information. An empty field means information was unavailable; it does not prove that the candidate lacks the experience. Use the evidence to decide what to verify in a conversation. Keep hiring decisions with the recruiting team and evaluate candidates against the same job-related criteria. If the search stops early, review the saved candidates or refine the next query. To stop a running Agentic Search, call POST /api/public/v1/agentic-search/{job_id}/cancel/ and poll until terminal. Work already performed may be billable within the credit cap; cancellation does not necessarily erase the charge. 4. Enrich shortlisted professional profiles Call POST /api/public/v1/enrich/professional-profile/ for candidates who need further review, using the professional URL from search. json { "professional_url": "https://www.linkedin.com/in/selected-candidate" } The placeholder above must be replaced with a real result. Professional Profile Enrichment endpoint-professional-profile returns structured current position, work history, education, and location when available. Preserve missing values rather than filling them with assumptions. Map the results into your ATS or candidate database through your application's integration. Store the source URL and search context alongside the candidate record. The Clodo search endpoint does not create an ATS record automatically. 5. Prepare recruiter outreach Use Email Enrichment endpoint-email to resolve an email from a selected professional URL. A lookup can return no match; do not create an outreach request until you have a usable address. Connect a sending account email-accounts , then submit the following structure to POST /api/public/v1/emails/ with actual candidate details. json { "to": { "email": "jordan@robotics.example", "first_name": "Jordan", "context": "Engineering leader with production robotics experience, reviewed by our recruiting team." }, "instructions": "Draft a concise recruiting introduction for our robotics engineering leadership role. Mention the relevant production experience without inventing accomplishments. Ask whether they would be open to learning about the role. Do not assume they are job hunting or available to relocate.", "review": "required" } Poll the email until pending_review , confirm the recipient and factual claims, and edit the draft if necessary. Only approve it when you want it queued for sending. Outreach Emails endpoint-outreach-emails documents draft edits, approval, cancellation, and reply state. Candidate sourcing questions How can I source candidates from a job description? Turn the description into professional requirements, preferences, and exclusions, then submit that persona to Agentic Search. Review returned evidence against the requirements and enrich selected profiles before outreach. Agentic Search queries support up to 2,000 characters; People Search and Deep Search support up to 500. Can I use People Search to build a talent pool? Yes. Describe the roles, industries, and geography you want. After a People Search completes, use search_id to retrieve additional non-overlapping pages. Deduplicate across separate searches in your own talent database. Does a match mean the candidate wants a new job? No. A professional match does not establish job-seeking status or interest. Ask the person directly, and keep outreach grounded in the experience you actually reviewed. Related workflows - Expert Discovery guide-experts : find practitioners for short research or advisory engagements. - GTM Lead Generation guide-gtm : adapt search and enrichment for sales prospecting. - Credits & Pricing credits : budget search caps, profile lookups, email lookups, and drafts separately. 

Recruiting & Candidate Sourcing

Find subject-matter experts for industry interviews, customer research, and advisory projects. Search by first-hand experience, review evidence, and prepare invitations.

 Expert Discovery with the Clodo API Find subject-matter experts for customer research, industry interviews, advisory projects, and market diligence. Describe the experience your project needs, research potential participants, and prepare an invitation that explains why their background is relevant. This workflow uses payments operations as an example: finding practitioners who have worked on clearing, settlement, and payment-network rules. You can adapt it to manufacturing, healthcare operations, energy, logistics, or another professional domain. 1. Define the expertise behind the research question Start with what you want to learn, then translate it into first-hand experience. “Fintech experts” is broad. “Payments operations leaders who have implemented clearing and settlement processes” identifies work that could make someone useful to interview. Research goal Useful persona --- --- Understand payment-network operations Practitioners with clearing, settlement, or scheme-rule implementation experience Assess warehouse automation adoption Operations leaders who have deployed robotics in distribution facilities Learn about enterprise software procurement Procurement or IT leaders who have evaluated and implemented relevant systems Specify geography or market context when it changes the expertise you need. Distinguish direct implementation experience from a general association with the industry. Describe the exact intersection of experiences your research needs. For example: “Someone who has run payments operations at both a bank and a fintech, worked on a clearing-system migration, and presented an implementation case study at an industry conference.” You can ask for technology experience, previous roles, published work, or participation in a named event together. State which experiences are essential and ask for evidence supporting them. 2. Find a researched expert shortlist Use Deep Search endpoint-deep-search for a focused professional query with a researched result set. Complete the API setup guide-use-cases , then send this JSON to POST /api/public/v1/deep-search/ with your x-api-key header. Use your own public HTTPS webhook receiver. json { "query": "Payments operations leaders in Singapore and Australia with experience implementing clearing and settlement processes or payment-network scheme rules. Prioritize hands-on implementation experience for an industry research interview.", "mode": "standard", "webhook_url": "https://your-app.example/webhooks/clodo" } Save the id from the accepted response. Receive deep_search.completed , or poll GET /api/public/v1/deep-search/{job_id}/ until terminal. Review each person's role, company, work history, and score . Treat the score as a search-fit signal rather than a certification of expertise. If the project needs a more detailed persona and supporting source evidence, choose Agentic Search endpoint-agentic-search instead. For a broad initial list of industry professionals, People Search endpoint-people-search may be sufficient. You do not need to run all three. 3. Use Agentic Search for a specific experience requirement For enrolled accounts, Agentic Search can research criteria that need more explanation. For example, distinguish someone who helped implement scheme rules from someone who merely works at a payments company. Send the following alternative request to POST /api/public/v1/agentic-search/ : json { "query": "Find payments practitioners in Singapore or Australia who have personally helped implement clearing and settlement processes or materially revise payment-network scheme rules. We want to interview people about implementation challenges. Prioritize first-hand operating experience and provide evidence of relevant work. Exclude sales-only roles.", "target_results": 15, "max_credits": 500, "webhook_url": "https://your-app.example/webhooks/clodo" } Inspect fit_summary , proof_points , and evidence in each returned person. Follow the cited sources as part of your review. Check partial and completion_reason when the job finishes, and record any unanswered expertise questions for screening. 4. Review experience and create your interview list Use Professional Profile Enrichment endpoint-professional-profile when you need structured work history for a selected person. Send their professional_url to POST /api/public/v1/enrich/professional-profile/ and inspect the returned positions and dates when available. In your own research database, record the question each person could help answer, the supporting experience, and what still needs confirmation. Compare expertise with the project requirements rather than assuming a senior title makes someone the best interviewee. Prepare a short screening question, such as: “Have you personally worked on implementing settlement operations, and which parts of that process can you discuss from experience?” Search identifies potential participants; the screening conversation confirms fit and willingness to participate. 5. Invite selected experts to a conversation Resolve contact details with Email Enrichment endpoint-email when needed. A resolved email does not establish availability or agreement to consult. Connect a sending account email-accounts , then create a reviewed invitation through POST /api/public/v1/emails/ . Use actual recipient details in place of this illustrative payload: json { "to": { "email": "sam@payments.example", "first_name": "Sam", "context": "Payments operations practitioner whose professional background is relevant to our clearing and settlement research." }, "instructions": "Write an invitation to a 30-minute industry research interview about clearing and settlement implementation. Explain why the recipient's professional background is relevant. Ask whether they have first-hand experience and are open to a conversation. Do not invent compensation, prior contact, or credentials.", "review": "required" } Poll the returned email resource until the draft is ready, review the wording, and approve it separately to queue sending. Use the returned reply state to update your own participant tracker. Scheduling an interview and managing an advisory engagement happen in your application or existing tools. Expert discovery questions How do I find subject-matter experts with an API? Describe the professional experience needed to answer your research question. Use Deep Search for a focused shortlist or Agentic Search for a detailed persona with supporting evidence. Review backgrounds and confirm expertise in a screening conversation. Can I find experts for customer discovery or market diligence? Yes. Search for practitioners with relevant operating experience, then review how their background relates to the research topic. The same workflow can identify potential interviewees, advisors, speakers, and industry participants. Does Clodo book interviews or provide an expert network? These API endpoints discover people, enrich profiles and contact details, and draft outreach. They do not book interviews, establish consulting terms, or guarantee that a person will participate. Related workflows - Recruiting & Candidate Sourcing guide-recruiting : adapt experience-based discovery to an open role. - GTM Lead Generation guide-gtm : find and qualify potential buyers. - Credits & Pricing credits : compare fixed search modes with an Agentic Search credit cap. 

Expert Discovery

Retrieve the status and results of asynchronous jobs.

 Async Polling Webhook delivery is the primary path for async search/enrichment results. Polling is the fallback when your webhook receiver is unavailable, when you need to fetch results out-of-band, or for debugging. Outreach Emails are the exception : they require no webhook secret and are designed to be polled — GET /api/public/v1/emails/{id}/ is the live resource itself not a job wrapper , so everything below about dj_ jobs doesn't apply to em_ emails. See Outreach Emails endpoint-outreach-emails . While drafting, poll every 5–10 seconds; drafts typically complete in 10–60 seconds. Polling URLs Use the job id returned by the original POST e.g. dj_a1b2c3d4... : Endpoint Polling URL --- --- People Search GET /api/public/v1/search/{job_id}/ Deep Search GET /api/public/v1/deep-search/{job_id}/ Agentic Search GET /api/public/v1/agentic-search/{job_id}/ Phone Enrichment GET /api/public/v1/enrich/phone/{job_id}/ Polling does not consume credits. Response shape json { "id": "dj_a1b2c3d4...", "status": "completed", "tier": "light", "created_at": "2026-05-02T06:39:06.013878Z", "started_at": "2026-05-02T06:39:06.079964Z", "completed_at": "2026-05-02T06:39:42.009761Z", "leads_returned": 25, "result": { ... }, "error": null } Field reference Field Type Notes --- --- --- id string dj_<32-hex job identifier. status string One of pending , running , completed , failed , cancelled . tier string One of light People Search , deep_standard , deep_extended Deep Search , agentic_search , phone_enrich . created_at ISO 8601 When the job was accepted. started_at ISO 8601 or null When the worker picked it up. completed_at ISO 8601 or null When the worker reached a terminal state. leads_returned int Final lead count. 0 until terminal. result object or null Populated only when status == "completed" . error object or null Populated only when status is failed or cancelled . result shape when completed The result object carries the same body the webhook would have delivered — minus the id / event_type wrapper. Schemas: - People Search: {results: ... , total_returned: int} - Deep Search: {results: ... , total_returned: int} - Agentic Search endpoint-agentic-search : {results: ... , total_returned: int, partial: bool, completion_reason: string} - Phone Enrichment: {phone: string null} Per-lead and per-field shapes are documented under Webhook Events webhook-events . error shape when failed or cancelled json {"code": "error", "message": "..."} Codes and messages are the same as the webhook .failed events — see Webhook Events webhook-events for the per-endpoint tables. Status lifecycle pending → running → completed → failed → cancelled pending is the initial state right after the POST returns 202. The worker transitions it to running when it picks up the job, then to one of the three terminal states. Agentic Search can also move directly from pending to cancelled . Cancelling a running Agentic Search normally finishes as completed with result.partial: true and result.completion_reason: "user_cancelled" . Always inspect the result's partial flag when processing completed Agentic Search jobs. Polling cadence Reasonable defaults: - People Search: poll every 5–10 seconds. Typical run completes in 15–30 seconds. - Phone Enrichment: poll every 10-15 seconds. Typical run completes in 1 minute. - Deep Search: poll every 30–60 seconds. Typical run completes in 10–15 minutes. - Agentic Search: poll every 30–60 seconds until terminal; duration depends on the research brief and budget. Polling endpoints have a high rate cap 3600/min so brief tight-loop polling will not throttle. Wrong job ID A GET against an unknown or malformed job_id returns 404 not_found in the canonical envelope. See also - Webhook Events webhook-events for the inner result shape and per-endpoint payloads. - Webhook Signing Secrets webhook-secrets and Webhook Signature Verification webhook-signing for the primary delivery path. - Idempotency idempotency for replaying without double-dispatch. 

Async Polling

 Idempotency Async POST endpoints accept an optional Idempotency-Key header. Same key on a retry returns the same job's current state instead of dispatching a new one. Supported on: - POST /api/public/v1/search/ - POST /api/public/v1/deep-search/ - POST /api/public/v1/agentic-search/ - POST /api/public/v1/enrich/phone/ Sync enrichment endpoints /enrich/email/ , /enrich/professional-url/ , /enrich/professional-profile/ do not support Idempotency-Key . The header is ignored if sent. Header Idempotency-Key: <your-key - Max 255 characters. Anything longer is truncated server-side. - Use a UUID, a request hash, or any client-generated unique value. - Reuse the same value across retries of a single logical request. - Generate a new value for each new logical request. Behavior - First call with a given key — creates a job, returns 202 with {id, status: "pending", created_at} . - Replay with same key — returns the same job's current state. No new dispatch, no new credit reservation. - Different key — creates a separate job, separate billing. Scope Keys are scoped to api_key, idempotency_key . Two different API keys on the same account using the same idempotency value get two separate jobs. Body mismatch on replay If you replay with the same Idempotency-Key but a different request body, we return the original job's state not the new request . Agentic Search is an exception: reusing a key with a different validated body returns 409 conflict , including changes to webhook_url . An unchanged replay returns the original job without another credit reservation. Use a new idempotency key for each new logical request, including requests to different endpoints. When to use - Network retries - Crash recovery Without an Idempotency-Key Each POST creates a new job and new credit reservation. Two identical POSTs back-to-back with no key produce two separate jobs and two separate charges. See also - People Search endpoint-people-search , Deep Search endpoint-deep-search , Agentic Search endpoint-agentic-search , Phone Enrichment endpoint-phone for the endpoint-specific request shapes. - Async Polling async-polling for fetching state of an existing job. 

Understand error responses and decide when to retry.

 Error Envelope Every error returned uses a canonical envelope. Parse it once and reuse across all endpoints. Shape json { "error": { "type": "invalid_request", "message": "Both first_name and last_name are required together — an orphan first or last name has no usable shape.", "doc_url": "", "request_id": "24a78305-e6a0-45e6-8645-59e3507b70f3", "retryable": false, "suggested_action": "Fix the highlighted field and retry.", "param": "non_field_errors" } } Field reference Field Type Notes --- --- --- type string Machine identifier. Stable across releases. message string Human-readable. Safe to display to end users. doc_url string Anchor link into these docs. May be empty. request_id string UUID per request. Also returned as X-Request-Id response header. retryable bool true if the same request would have a chance of succeeding on retry. suggested_action string Concrete next step. param string Optional. Set on validation errors to identify the failing field e.g. "domain" , "non_field_errors" . Codes HTTP error.type When Retryable --- --- --- --- 400 invalid_request Request body or query params failed validation no 402 insufficient_credits Account balance insufficient for the call no 403 permission_denied Key valid but account not eligible for the public API no 404 not_found Resource does not exist no 409 conflict Request conflicts with existing state no 429 rate_limited Agentic Search concurrency limit exceeded gateway rate limits use the response below yes 500 internal_error Server-side error. no 502 upstream_error Upstream error. yes 503 service_unavailable Service temporarily unavailable. yes Two error responses do NOT use this envelope 403 Forbidden — x-api-key missing, revoked, or not recognized: json {"message": "Forbidden"} 429 — rate limit exceeded: HTTP/1.1 429 Too Many Requests {"message":"Too Many Requests"} See also - Authentication authentication for 403 Forbidden semantics. - Per-endpoint pages for endpoint-specific rate limits. 

Handling Errors

Plan your request throughput and handle rate limiting.

 Rate Limits Rate limits are enforced per endpoint . See each endpoint's reference page for its sustained rate and burst. How limits are enforced Each endpoint has two caps: - Sustained rate — requests per minute, smoothed over time. - Burst — short-window allowance over the sustained rate. Exceeding either returns 429 . Agentic Search endpoint-agentic-search also limits pending/running jobs to 2 per API key and 3 per account. Its concurrency rejection uses the API error envelope with error.type: "rate_limited" ; the gateway rate-limit response below uses a different format. 429 response shape HTTP/1.1 429 Too Many Requests {"message":"Too Many Requests"} We do not return a Retry-After header and recommend jittered exponential backoff. Catch-all fallback Requests to paths not in the explicit endpoint list typos, missing trailing slash, deprecated paths fall through to a catch-all with a much tighter limit: Limit Value --- --- Sustained rate 1 request / minute Burst 2 requests /enrich/email no slash hits the catch-all; /enrich/email/ hits the explicit endpoint. See also - Email Enrichment endpoint-email - Phone Enrichment endpoint-phone - Professional URL Enrichment endpoint-professional-url - Professional Profile Enrichment endpoint-professional-profile - People Search endpoint-people-search - Deep Search endpoint-deep-search - Agentic Search endpoint-agentic-search 

Idempotency

#Header

#Behavior

#Scope

#Body mismatch on replay

#When to use

#Without an Idempotency-Key

#See also