Browse published operations

An overview of paid and organic keywords for a domain, with estimation of budgets and CPC included.

See which Facebook or Instagram ads a company is running by providing their website or Facebook page as input.

Find ads in the ad library based on a keyword.

See which Google ads a company is running by providing their website as input.

Retrieve all Google shopping ads for a specified company by providing their domain. This differs from the standard Google endpoint, which reports search, image, and video ads on Google's platform.

Retrieve results from a Google Shopping search using a request ID obtained from the initial Google Shopping endpoint call. The service is free to use.

See which LinkedIn ads a company is running by providing their website as input.

This endpoint runs a search in the LinkedIn Ad Library with the keyword and, optionally, the filters you provide.

Get all the details for an ad on TikTok using the ad id as the starting point.

Do an ads search on the platform using a keyword.

Searches 70+ million enriched, verified and active company profiles using account filters — ideal for building target-account (ABM) lists. Includes Lookalike Search: pass up to 5 domains or LinkedIn company URLs in lookalikeDomains to find similar companies. 💳 Credit Usage: 0.1 credits per returned result (Lookalike Search included). E.g. size: 100 with 100 matches = 10 credits. 💡 Tip: Each result includes a company id (UUID). Use these IDs in the People Search filter contact.company to find people working at exactly these companies — more precise than matching by name. Company Identification: Name or Domain, Domain, LinkedIn, URL, Name, Social Media Link, Phone Number Company Details: Industry, Location, Products and Services, Social Media, Company Type, Founded Year, Operation Language, Geo Location Financial & Size: Employee Size, Annual Revenue, Funding Advanced Filters: Keywords, Employee by Department, Headcount Growth, Technologies, NAICS 🔗 Related Endpoints: • People Search — find the people at these companies (0.5/result) • Export People with Email — bulk-export those people with verified emails • Lists — exclude accounts you already work (free) 📥 Filter value lists: industries.csv (919) · technologies.csv (16,000+) · departments-and-functions.csv (592). All other allowed values are listed inline in the field

Searches 400+ million enriched, verified and active person profiles using contact and/or account (company) filters — the starting point for building outbound lead lists. Each result contains the full person profile and their company data. Results do not include emails or phone numbers — enrich them with the endpoints listed under Related Endpoints below. 💳 Credit Usage: 0.5 credits per returned result (e.g. size: 25 with 25 matches = 12.5 credits). You only pay for results actually returned. ⚠️ Track ID: The response includes a trackId. Pass it to Find Emails by Track ID to get verified emails for this exact result set. The trackId can be used once and expires 6 hours after this response — submit it promptly. Contact Filters: Profile: Full Name, Social Media Link, LinkedIn, Location, Language Skills, Profile Badge, LinkedIn Follower & Connection Count Professional: Current & Past Company, Seniority, Department and Function, Experience (Current/Previous Title, Duration) Skills & Education: Skills, Certifications, Education (School, Degree, Field of Study, Date), Keywords, Social Media Account Filters: Company Identification: Name or Domain, Domain, LinkedIn, URL, Name, Social Media Link, Phone Number Company Details: Industry, Location, Products and Services, Social Media, Company Type, Founded Year, Operation Language, Geo Location Financial & Size: Employee Size, Annual Revenue, Funding Advanced Filters: Keywords, Employee by Department, Headcount Growth

Analyzes the skills, headline, and summary of a profile and returns a complete personality profile with ready-to-apply outreach guidance: DISC and OCEAN (Big Five) scores with an archetype label, plus selling and hiring blocks containing field-by-field email advice (subject, salutation, tone, length, bullet points, closing line), whatToSay / whatToAvoid talking points, and key decision traits (risk tolerance, decision speed, decision drivers). Pipe these into your email-generation prompts to personalize outreach at scale. 💳 Credit Usage: 4 credits per request. 🔗 Related Endpoints: • People Search — find the people (and their LinkedIn URLs) first (0.5/result) • Export Single Person with Email — their BounceBan-verified email (1 credit) • Reverse People Lookup — identify a person from an email/phone before analyzing them (0.5/request) The header Content-Type: application/json must be included in the request.

Triggers email finding for a previous People Search result set, using the trackId from the People Search API response. This is step 2 of the two-step list-building flow (search → find emails). 💳 Credit Usage: 1 credit per found valid email (0.5 for the people data + 0.5 for the real-time BounceBan-verified email) — and 0 credits for people where no valid email is found. Example: 25 people submitted, 18 valid emails found = 18 × 1 = 18 credits; the other 7 cost nothing. Unlike Export People with Email (0.5 per person regardless + 0.5 per found email), here you only pay for delivered emails. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan — safe to send, no re-verification needed. ⚠️ Important: Each trackId can only be used once. After the first request, the track ID is consumed and subsequent requests with the same ID will fail. The trackId also expires 6 hours after the original People Search response. Provide a webhook URL to receive asynchronous notifications when email finding completes. When the job completes, we POST the result to your webhook URL. Auto-refund window: A submission that is charged but never delivered is auto-refunded after up to 10 hours. Rate limits (per token): If a token already has 500 requests in flight, a further submit is rejected with 400 too many pending requests — drain below the cap (or wait for completions), then retry. 10 jobs are processed at once. Limits are independent per service and per token. The response includes statistics</cod

💳 Credit Usage: Free — fetching results never costs credits (you already paid at submission). Paginate with page (zero-based) and size (1–100, default 10). Export People Results: returns 409 while the export is still in progress — poll Statistics until state: DONE. Email Finder Results: readable while running; unfinished items show state: PROCESSING. A fully refunded (STUCK_*) trackId returns 403 — resubmit it. Returns paginated email finder results for a given trackId. The trackId is from the Find Emails by Track ID response. Use the page and size query parameters to paginate through results. Each item includes a refId, state, input (firstname, lastname, domain), and output (email results with verification details). ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan. Fully-refunded submissions: if the submission was auto-refunded for non-delivery (STUCKHARDREFUND / STUCKSOFTREFUND), this endpoint returns 403 Forbidden with an actionable message — the results are not retrievable, so submit the request again. If it was refunded because there was no deliverable result (SEARCHNORESULTS / NOEMAILSFOUND), it returns 200 with an empty page.

💳 Credit Usage: Free — polling never costs credits (you already paid at submission). Use this to track state and statistics until the job is DONE (Email Finder: total, found · Export: total, success, failed, found). A fully refunded (STUCK_*) trackId returns 403 with the refund reason — resubmit it. Returns email-finding statistics for a given trackId. The trackId is from the Find Emails by Track ID endpoint. Use this endpoint to poll progress (e.g. statistics.total, statistics.found) and state until the job completes. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan. Fully-refunded submissions: if the submission was auto-refunded for non-delivery (STUCKHARDREFUND / STUCKSOFTREFUND), this endpoint returns 403 Forbidden with an actionable message — the results are not retrievable, so submit the request again. If it was refunded because there was no deliverable result (SEARCHNORESULTS / NOEMAILSFOUND), it returns 200 with zeroed statistics. The header Content-Type: application/json must be included in the request.

Search and email finding in a single asynchronous call: exports up to 10,000 people matching the same account and contact filters as the People Search API, with real-time verified emails included. The recommended endpoint for bulk outbound list building. 💳 Credit Usage: 0.5 credits per exported person (charged for everyone, whether or not an email is found) + 0.5 credits per found valid email. Example: 1,000 people with 700 emails found = 1,000 × 0.5 + 700 × 0.5 = 850 credits. You only pay the email portion for emails verified in real time by BounceBan. Note: this differs from Find Emails by Track ID, which charges only for found emails. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan — safe to send, no re-verification needed. Limit: size maximum is 10000. Exceeding this limit returns 400 Bad Request - pagination limit exceeded. Results are processed asynchronously; provide a webhook URL to receive a POST when the export completes, or poll Export People Statistics. Auto-refund window: A submission that is charged but never delivered is auto-refunded after up to 10 hours. Rate limits (per token): If a token already has 500 requests in flight, a further submit is rejected with 400 too many pending requests — drain below the cap (or wait for completions), then retry. 10 jobs are processed at once. Limits are independent per service and per token. The response returns a trackId and st

💳 Credit Usage: Free — fetching results never costs credits (you already paid at submission). Paginate with page (zero-based) and size (1–100, default 10). Export People Results: returns 409 while the export is still in progress — poll Statistics until state: DONE. Email Finder Results: readable while running; unfinished items show state: PROCESSING. A fully refunded (STUCK_*) trackId returns 403 — resubmit it. Returns paginated export results for a given trackId. The trackId is from the Export People with Email response. Use the page and size query parameters to paginate through results. Returns full people data with email, company, and profile information. When email finding is still in progress, items may have email.state PROCESSING and an empty email.output array. When complete, email.state is DONE and email.output contains the found email(s). ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan. Returns 409 Conflict when the track ID is still in progress. Fully-refunded submissions: if the submission was auto-refunded for non-delivery (STUCKHARDREFUND / STUCKSOFTREFUND), this endpoint returns 403 Forbidden with an actionable message — the results are not retrievable, so submit the request again. If it was refunded because there was no deliverable result (SEARCHNORESULTS / NOEMAILSFOUND), it returns 200 with an empty page.

V2 of Export Single Person with Email — Clay-compatible. Behaves exactly like /v1/people/export/single, except that when the person is not found or has no email it returns HTTP 200 with an envelope { status, error, data } and data: null (instead of the v1 404). Lookup uses id first, then url. You must provide either id or url; if both are empty the API returns 400 Bad Request. ✨ Recommended for Clay: this V2 endpoint gracefully returns HTTP 200 for both the found and not-found cases, which is what Clay expects when enriching leads. 💳 Credit Usage: 1 credit per successful call (0.5 Full Profile Enrichment + 0.5 real-time BounceBan-validated email), and 0 credits if we do not find a valid email. The charged amount is also returned in the X-Credit response header (e.g. -1.0), present only on the charged (email found) response. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan — safe to send, no re-verification needed. Real errors (400 / 402 / 500) and gateway errors (401 / 429) are returned as-is with their real HTTP status — they are not wrapped. 🔗 Related Endpoints: • Export Single Person with Email (V1) — same pricing, 404 on not-found • Mobile Phone Finder (V2 – Clay Compatible) — add mobile numbers in the same Clay table • Export People with Email — bulk instead of row-by-row The header

Exports a single person with email finding in real time — built for row-by-row enrichment (CRM records, spreadsheets, Clay tables). Returns the full person profile plus a verified email. You must provide either id (the person's AI-Ark ID from a previous People Search) or url (the person's LinkedIn profile URL). At least one of the two fields is required; if both are empty, the API returns 400 Bad Request. If no email is found, the API returns 404 Not Found. 💳 Credit Usage: 1 credit per successful call (0.5 for the Full Profile Enrichment and 0.5 for the real-time BounceBan-validated email) — and 0 credits if we do not find a valid email. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan — safe to send, no re-verification needed. Using Clay? The Clay-compatible /v2/people/export/single endpoint is recommended: it returns HTTP 200 for found and not-found cases, which is what Clay expects. 🔗 Related Endpoints: • Export Single Person with Email (V2 – Clay Compatible) • Export People with Email — bulk export up to 10,000 people at once • Mobile Phone Finder — add a mobile number for the same person (5 credits/result) • Reverse People Lookup — start from an email or phone instead The header Content-Type: application/json must be included in the request.

💳 Credit Usage: Free — polling never costs credits (you already paid at submission). Use this to track state and statistics until the job is DONE (Email Finder: total, found · Export: total, success, failed, found). A fully refunded (STUCK_*) trackId returns 403 with the refund reason — resubmit it. Returns export statistics for a given trackId. The trackId is from the Export People with Email endpoint. Use this endpoint to poll progress (e.g. statistics.total, statistics.found) and state until the job completes. ✅ Email Verification: All emails (SMTP & CATCH_ALL) returned by the API are verified in real time by BounceBan. Fully-refunded submissions: if the submission was auto-refunded for non-delivery (STUCKHARDREFUND / STUCKSOFTREFUND), this endpoint returns 403 Forbidden with an actionable message — the results are not retrievable, so submit the request again. If it was refunded because there was no deliverable result (SEARCHNORESULTS / NOEMAILSFOUND), it returns 200 with zeroed statistics. The header Content-Type: application/json must be included in the request.

V2 of the Mobile Phone Finder — Clay-compatible. Behaves exactly like /v1/people/mobile-phone-finder, except that when no person or phone is found it returns HTTP 200 with an envelope { status, error, data } and data: null (instead of the v1 404). Search Options: • LinkedIn Search: Provide only the linkedin URL (no need for domain/name) • Domain + Name Search: Provide both domain and name (no need for LinkedIn) ✨ Recommended for Clay: gracefully returns HTTP 200 for both found and not-found cases, which is what Clay expects when enriching leads. Learn how to use this API inside Clay: Watch Video 💳 Credit Usage: 5 credits per result (charged when a phone is found), 0 otherwise. The charged amount is also returned in the X-Credit response header (e.g. -5.0), present only on the charged (phone found) response. Real errors (400 / 402 / 500) and gateway errors (401 / 429) are returned as-is with their real HTTP status — they are not wrapped. 🔗 Related Endpoints: • Mobile Phone Finder (V1) — same pricing, 404 on not-found • Export Single Person with Email (V2 – Clay Compatible) — verified email in the same Clay flow (1 credit) Rate limits: 5 requests per second by default. Customers using more than 450,000 credits per month can get custom rate limits — contact us a

Finds mobile phone numbers for a person — use it to add direct-dial numbers to your outbound lists. Search Options: • LinkedIn Search: Provide only the linkedin URL (no need for domain/name) • Domain + Name Search: Provide both domain and name (no need for LinkedIn) 💳 Credit Usage: 5 credits per result (charged when a phone number is found). If nothing is found, the API returns 404 and no credits are charged. Using Clay? The Clay-compatible /v2/people/mobile-phone-finder endpoint is recommended for enriching leads inside Clay. 🔗 Related Endpoints: • Mobile Phone Finder (V2 – Clay Compatible) • Export Single Person with Email — verified email for the same person (1 credit) • People Search — find the people (and their LinkedIn URLs) first • Reverse People Lookup — identify who a phone number belongs to Rate limits: 5 requests per second by default. Customers using more than 450,000 credits per month can get custom rate limits — contact us at support@ai-ark.com or via the in-app chat. The header Content-Type: application/json must be included in the request.

Run any People Search for a flat 1 credit per page and see who matches before spending per-result credits — the recommended first call of every list-building workflow. Validate your filters, check totalElements, inspect match quality, then send the exact same filter body to People Search or Export People with Email. 💳 Credit Usage: 1 credit per page — whether the page holds 25 or 100 results. A full page of 100 previews costs 1 credit; the same page on People Search costs 50 credits (0.5 × 100). Up to 100 results per page (default 25), max 10,000 results per search. Each preview result contains: • The person's real AI Ark id — accepted directly by Export Single Person with Email, no full search needed • First name, masked last name (e.g. Ba*e), title, headline, location, department/functions/seniority, position history with dates • The full company object: id, name, description, industry, headcount, revenue, HQ + office locations, keywords, NAICS/SIC • has* availability flags (hasmobile, hasskills, haseducations, …) — see what data exists before paying to enrich; has_mobile: true means a Mobile Phone Finder call (5 credits) has something to find Not included: unmasked last names, emails, phone numbers, and no trackId (email finding chains off People Search only). Body: identical to People Search — all account, contact and lists filters plus page and size. 🔗 <str

Identifies a person from an email address: if a person with that email exists in the AI Ark database, you get back their full profile (name, title, company, LinkedIn, and more) — similar to a People Search result, excluding valid email and mobile phone. Useful for enriching inbound leads, newsletter signups, form fills, and CRM records where you only have an email. A 404 means no profile with this email was found. The search parameter contains the email address to look up and is the only request field. 💳 Credit Usage: 0.5 credits per request. 🔗 Related Endpoints: • Export Single Person with Email — get a BounceBan-verified email for the identified person (1 credit) • Mobile Phone Finder — get their mobile number (5 credits/result) • Personality Analysis — personalize your outreach to this person (4 credits) 📥 Filter value lists: industries.csv (919) · technologies.csv (16,000+) · departments-and-functions.csv (592). All other allowed values are listed inline in the field descriptions below. Learn how to use this API to enrich leads inside Clay: Watch Video Rate limits: 5 requests per second by default. Customers using more than 450,000 credits per month can get custom rate limits — contact us at support@ai-ark.com or via the in-app chat. The hea

Fetch a company's employee review data including overall ratings, dimension-level ratings (culture, work-life balance, senior management and more) and individual employee reviews. Results are returned in the API response.

Retrieve structured company data — firmographics, funding, headcount, web traffic, business model, and financials. Specify which sections you need; only requested sections are returned.

Fetches a company's LinkedIn-sourced headcount data, including total employee count, historical growth trends, and a functional breakdown of the workforce.

Fetches live LinkedIn and Indeed job listings for a company, including title, location, description, compensation, experience level, and extracted key skills.

Fetches social media posts published by a company, including content type, text, publish date, paid/repost flags, an AI classification, and detailed engagement metrics.

Fetches a company's product catalog and detailed reviews per product including ratings, star distribution, pros, cons, pricing and individual review content. Results are returned in the API response.

Resolve a company name, website domain, or UUID into a structured company record. This is a free endpoint — it does not consume API credits.

Fetches a company's website traffic signals, including engagement metrics, historical monthly visit estimates, and a breakdown of traffic by acquisition source.

Resolves an industry from a free-text query and returns matching industry codes ranked by similarity. This is a free endpoint — it does not consume API credits.

Converts natural language queries into structured filter conditions for the List Generation API.

Build targeted company lists using structured filters or natural language queries, with optional data enrichment.

Fetch enriched news articles for a company or industry. Each article includes AI-generated summary, full text, sentiment, event tags, industry classifications, and resolved company mentions. Results are returned synchronously — no polling required.

Returns the current status of an asynchronous request, such as a company addition request, by its request ID. This is a free endpoint — it does not consume API credits.

Get the default build for an Actor. Use the optional waitForFinish parameter to synchronously wait for the build to finish. This avoids the need for periodic polling when waiting for the build to complete. This endpoint does not require an authentication token. Instead, calls are authenticated using the Actor's unique ID. However, if you access the endpoint without a token, certain attributes (e.g., usageUsd and usageTotalUsd) will be hidden.

Gets the list of builds of a specific Actor. The response is a JSON with the list of objects, where each object contains basic information about a single build. The endpoint supports pagination using the limit and offset parameters and it will not return more than 1000 records. By default, the records are sorted by the startedAt field in ascending order, therefore you can use pagination to incrementally fetch all builds while new ones are still being started. To sort the records in descend

Get the OpenAPI definition for Actor builds. Two similar endpoints are available: - First endpoint: Requires both actorId and buildId. Use default as the buildId to get the OpenAPI schema for the default Actor build. - Second endpoint: Requires only buildId. Get the OpenAPI definition for a specific Actor build. To fetch the default Actor build, simply pass default as the buildId. Authentication is based on t

Builds an Actor. The response is the build object as returned by the Get build endpoint.

Deletes an Actor with the specified ID.

Gets the list of all Actors that the user created or used. The response is a list of objects, where each object contains a basic information about a single Actor. To only get Actors created by the user, add the my=1 query parameter. The endpoint supports pagination using the limit and offset parameters and it will not return more than 1000 records. By default, the records are sorted by the createdAt field in ascending order, therefore you can use pagination to incrementally fetch all Act

Gets an object that contains all the details about the Actor with the specified ID.

Creates an Actor with the settings specified in an Actor object passed as JSON in the POST payload. Returns the created Actor object. In the HTTP request, set the Content-Type header to application/json. ### Define a source code version An Actor must specify at least one version of the source code. For details, see Actor versions. ### Publish an Actor To make your Actor public, use the [Update Actor](/a

Updates an Actor's settings with the values specified in an Actor object passed as JSON in the POST payload. Only the fields specified in the request body are updated. Returns the full Actor object, the same as the Get Actor endpoint. In the HTTP request, set the Content-Type header to application/json. ### Authentication To provide the authentication token, we recommend using the request's Authorization header, rather than the URL. For details, see [Authentication

Runs a specific Actor and returns a key-value store record. The response contains the record stored under the OUTPUT key in the run's default key-value store. This is a legacy approach that has been replaced by the Actor output object; the record may not exist, in which case the response contains no data. Use the outputRecordKey query parameter to return a different record. The run m

Runs a specific Actor and returns its dataset items. The run must finish in 300<!-- MAXACTORJOBSYNCWAIT_SECS --> seconds otherwise the API endpoint returns a timeout error. The Actor is not passed any input. It allows to send all possible options in parameters from Get Dataset Items API endpoint. Beware that it might be impossible to maintain an idle HTTP connection for a long period of time, due to client timeout or network conditions. Make

Runs a specific Actor and returns its dataset items. The POST payload including its Content-Type header is passed as INPUT to the Actor (usually application/json). The HTTP response contains the Actors dataset items, while the format of items depends on specifying dataset items' format parameter. You can send all the same options in parameters as the Get Dataset Items API endpoint. The Actor is started with the default options; you can ov

Runs a specific Actor and returns a key-value store record. The POST payload including its Content-Type header is passed as INPUT to the Actor (usually application/json). The response contains the record stored under the OUTPUT key in the run's default key-value store. This is a legacy approach that has been replaced by the Actor output object; Actors aren't required t

Gets the list of runs of a specific Actor. The response is a list of objects, where each object contains basic information about a single Actor run. The endpoint supports pagination using the limit and offset parameters and it will not return more than 1000 array elements. By default, the records are sorted by the startedAt field in ascending order, therefore you can use pagination to incrementally fetch all records while new ones are still being created. To sort the records in descending

This is not a single endpoint, but an entire group of endpoints that lets you to retrieve and manage the last run of given Actor or any of its default storages. All the endpoints require an authentication token. The base path represents the last Actor run object is: /v2/actors/{actorId}/runs/last{?token,status,origin} Using the status query parameter you can ensure to only get a run with a certain status (e.g. status=SUCCEEDED). Similarly, the origin query parameter filters runs by the m

Deletes the default dataset associated with the last Actor run. This endpoint is a shortcut for getting the last run's defaultDatasetId and then using the Delete dataset endpoint.

Returns the default dataset associated with the last Actor run. This endpoint is a shortcut for getting the last run's defaultDatasetId and then using the Get dataset endpoint.

Returns data stored in the default dataset of the last Actor run in the desired format. This endpoint is a shortcut that resolves the last run's defaultDatasetId and proxies to the Get dataset items endpoint.

Appends an item or an array of items to the end of the last Actor run's default dataset. This endpoint is a shortcut that resolves the last run's defaultDatasetId and proxies to the Store items endpoint. To save bandwidth and speed up your upload, you can send the request payload compressed and set the Content-Encoding header accordingly. Below is a list of supported Content-Encoding types. Brotli: Content-Encoding: br Gzip: Content-Encoding: gzip * De

Updates the default dataset associated with the last Actor run. This endpoint is a shortcut for getting the last run's defaultDatasetId and then using the Update dataset endpoint.

Returns statistics for the last Actor run's default dataset. This endpoint is a shortcut that resolves the last run's defaultDatasetId and proxies to the Get dataset statistics endpoint.

Deletes the last Actor run key-value store. This endpoint is a shortcut for getting the run's defaultKeyValueStoreId and then using the Delete store endpoint.