Skip to content

Updates

What's changing

Retirements, breaking changes and maintenance that affect your integration now or soon, with what to do, then every release of the API.

Notices updated . What already shipped is in the changelog below.

In effect

  1. Breaking change
    Renamed

    Bulk jobs are now tasks, at /v2/bulk and /v2/tasks

    Bulk starts at POST /v2/bulk, and its jobs, now called tasks, are read under /v2/tasks. The request field name is taskName, the response holds data.task, the webhook sends task.updated with taskId, and the error codes say TASK. A signal import into a List (POST /v2/lists/{listId}/intake/signals) also answers data.task.

    • POST /v2/enrich/persons/bulkPOST /v2/bulk
    • GET /v2/jobsGET /v2/tasks
    • GET /v2/jobs/{jobId}GET /v2/tasks/{taskId}
    • GET /v2/jobs/{jobId}/resultsGET /v2/tasks/{taskId}/results
    • GET /v2/jobs/{jobId}/rowsGET /v2/tasks/{taskId}/rows
    • POST /v2/jobs/{jobId}/cancelPOST /v2/tasks/{taskId}/cancel
    The API answers
    404 with the error code NOT_FOUND: the former routes have no alias. A task started before the change keeps its id, read at /v2/tasks/{taskId}. The call costs no credits.
    What to do
    Call the new routes, send taskName in place of name, read data.task, and listen for task.updated. The SDKs 0.4.0 and the CLI 0.8.0 make the change for you, and the MCP server lists its new tools at the next connection. Live fetches and contact lookups keep their routes and fields.
    Every renamed field
  2. Breaking change
    In effect

    Search autocomplete refuses a one-character skills query

    Search autocomplete (POST /v2/search/people/auto-completion-fields, and its former path /v2/search/people/auto-suggestion) now has a shortest query for each field. For skills it is two characters: a one-character query, accepted until now, is refused. industry, countryCode and area still answer from the first character.

    • POST /v2/search/people/auto-completion-fields
    The API answers
    422 with the error code VALIDATION_ERROR and a message that names the minimum: "query" must be at least 2 characters for "skills", received 1. The call costs no credits.
    What to do
    Send a skills query from its second character. A search box that calls the endpoint on each keystroke waits for that character before its first call.
    The shortest query of each field

In progress

No maintenance in progress.

Upcoming

  1. Maintenance
    Deprecated
    Paused
    Back
    in a few weeks

    The former company search endpoint goes into maintenance on November 9

    POST /v2/search/companies and its available-fields route keep working until November 8, 2026, then pause for a few weeks from November 9. Company search at /v2/search/organizations does the same job now.

    • POST /v2/search/companiesPOST /v2/search/organizations
    • GET /v2/search/companies/available-fieldsGET /v2/search/organizations/available-fields
    The API answers
    From November 9, 2026, 410 with the error code ENDPOINT_GONE. error.details.alternative names the endpoint to call and error.details.docs links the migration guide. The call costs no credits.
    What to do
    Move to /v2/search/organizations before November 9, 2026. /v2/search/companies comes back in a format compatible with it, so there is no second migration.
    Migration guide

API changelog

Every change to the Reverse Contact API: new endpoints, improvements, fixes and deprecations.

Last updated

  1. v2.46.4

    Search autocomplete moves to /v2/search/people/auto-completion-fields

    Search autocomplete for your product now answers at POST /v2/search/people/auto-completion-fields, next to the available fields of People search, and returns its values under completions.

    🚀 Improvements

    • Search autocomplete is now POST /v2/search/people/auto-completion-fields, with the same request (field, query, limit), the same limits and still 0 credits (see Search autocomplete for your product)
    • Its response holds the values in data.completions, under the object search_auto_completions: each one keeps value, to send in the filter, and label, to show
    • The TypeScript and Python SDKs 0.5.0 add search.peopleAutoCompletionFields (search.people_auto_completion_fields in Python), which calls the new path and returns data.completions

    ⚠️ Deprecated

    • POST /v2/search/people/auto-suggestion, the former path, still answers the same request with the values in data.suggestions under the object search_auto_suggestions: move to the new path when you next touch your integration
    • In the SDKs, search.peopleAutoSuggestion (search.people_auto_suggestion in Python) is deprecated and keeps working on the former path
  2. v2.46.3

    Search autocomplete for job titles, companies, schools, languages, cities and regions

    Search autocomplete for your product now completes job titles, company names, schools, languages, cities and regions as your users type, besides skills, industries, countries and areas.

    🚀 Improvements

    • POST /v2/search/people/auto-suggestion takes eight more values as field: positionTitle, currentPositionTitle, companyName, currentCompanyName, schoolName, languages, city and region, still for 0 credits (see Search autocomplete for your product)
    • Each of them returns the values found on profiles that start with the query, most common first, in lowercase: sof returns software engineer for positionTitle, sa returns sao paulo for city
    • Each field is named after the People search filter its value goes back into: city and region complete location.city and location.region
    • A query now has a shortest length per field: 3 characters for job titles and company names, 2 for schools, skills, languages, cities and regions, 1 for industries, countries and areas. A shorter query returns 422 VALIDATION_ERROR with the minimum in its message
    • A skills query of one character, accepted until now, returns that 422: start skills at two characters
    • The TypeScript and Python SDKs 0.4.2 accept the new field values
  3. v2.46.2

    Polling results and request logs kept 7 days

    Async results and request logs now stay available 7 days by default instead of 72 hours.

    🚀 Improvements

    • Results from GET /v2/webhooks/:webhookId are retained 7 days by default, counted from the moment the task is created (previously 72 hours), before the endpoint returns 410 with WEBHOOK_EVENT_EXPIRED (see Result retention)
    • The request logs and webhook events of your dashboard are kept 7 days as well (previously 72 hours)
  4. v2.46.1

    Check profiles in bulk, Lists export as NDJSON, search autocomplete for industries, countries and areas

    Bulk can now check a whole file of profiles for free, a List now exports as NDJSON for your code, with each member's stored profile and company, and search autocomplete for your product completes industries, countries and areas besides skills.

    🚀 Improvements

    • POST /v2/bulk takes families: ["check"], which runs the free Person profile status check on each item that has a Social profile URL, for 0 credits per item (see Check profiles in bulk)
    • check runs alone: sent with another family or with live: true, the request returns 422 VALIDATION_ERROR, and like Person profile status, a check task needs at least 2 credits in your workspace (402 NO_CREDITS below that)
    • In the results of a check task, an item reads found or not_found, an item without a Social profile URL reads skipped, the CSV adds rc_exists, rc_last_update, rc_experience_count, rc_skill_count and rc_school_count, and each NDJSON line carries check (see Download task results)
    • The TypeScript and Python SDKs 0.4.1 accept families: ["check"] and type the check of TaskResultLine (TaskResultLineCheck in Python), and the CLI 0.8.1 takes rc bulk create --families check
    • GET /v2/lists/{listId}/export takes format=ndjson, like the results of a Bulk task: one JSON object per line, the row of List members, then person (the stored profile, in the shape of Enrich profile), company (in the shape of Enrich company), isFullProfile and customColumns, the columns your team added in the app, still for 0 credits (see Export a List)
    • The CSV stays the default, and its fields parameter, which chooses the columns and their order, is now documented (see Choose the columns). With format=ndjson, fields returns 400 LIST_INVALID_INPUT
    • The MCP tool export_list takes format: "ndjson", and the CLI 0.8.1 adds --format and --fields to rc lists export (see CLI)
    • The TypeScript and Python SDKs 0.4.1 take format and fields in lists.export and export the ListExportLine type. In TypeScript, the request options of lists.export come third, as in bulk.results, and the former call keeps working
    • POST /v2/search/people/auto-suggestion takes industry, countryCode and area as field besides skills, still for 0 credits (see Search autocomplete for your product)
    • A country is found by its ISO code or its English name: fr and fran both return France first, and UK returns the United Kingdom
    • An area is found by its id, its acronym or its name: dach returns MC:DACH
    • Each suggestion now carries label, the text to show, next to value, the text to send in the filter: { "value": "FR", "label": "France" }. A skill's label is the skill itself
    • The TypeScript and Python SDKs 0.4.1 accept the new field values and type label
    • The OpenAPI spec has one public address, https://docs.reversecontact.com/openapi.json or /openapi.yaml, which gives each operation its docs page (externalDocs), credit cost (x-credits), mode (x-mode) and status (x-availability) (see Agent brief)
    • https://api.reversecontact.com/openapi.json now redirects there (301) and serves the same spec, and https://api.reversecontact.com/openapi.yaml redirects to the YAML copy (previously 404)

    ⚠️ Deprecated

    • In the TypeScript SDK, lists.export(listId, options) is deprecated: pass the request options third, lists.export(listId, {}, options)
  5. v2.46.0

    Jobs become tasks, with Bulk at /v2/bulk and /v2/tasks

    What the API called jobs are now tasks, so they no longer read as job positions: Bulk starts at POST /v2/bulk, its tasks live under /v2/tasks, and live fetches, contact lookups and List imports say task too.

    🚀 Improvements

    • Live fetches and contact lookups are now called tasks in the API messages and the docs, while their routes and fields stay as they are (data.webhookId, pollUrl, GET /v2/webhooks/{webhookId})
    • Error messages and result rows say task in place of job, for example Skipped because the task was cancelled
    • The TypeScript and Python SDKs 0.4.0 add client.tasks, Task, TaskFailedError and TaskTimeoutError to follow live fetches and contact lookups (see client.tasks)

    💥 Breaking Changes

    • POST /v2/enrich/persons/bulk is now POST /v2/bulk (see Bulk)
    • The routes under /v2/jobs move to /v2/tasks: GET /v2/tasks, GET /v2/tasks/{taskId}, GET /v2/tasks/{taskId}/results, GET /v2/tasks/{taskId}/rows and POST /v2/tasks/{taskId}/cancel (see Get a task)
    • The former routes answer 404 with the error code NOT_FOUND, with no alias, and a task started before the change keeps its id at /v2/tasks/{taskId}
    • The request field name is now taskName, and the task object returns taskName in place of name
    • The 202 response of Bulk holds the task in data.task (previously data.job), and its links.self and links.results point to /v2/tasks/{taskId}
    • The 202 response of POST /v2/lists/{listId}/intake/signals holds the same record in data.task (previously data.job)
    • The webhook event is now task.updated (previously job.updated), in the body and in the x-reversecontact-event header, with taskId in place of jobId and pollUrl and resultsUrl under /v2/tasks
    • Error codes say TASK in place of JOB, for example TASK_NOT_FOUND, TASK_RESULTS_NOT_READY, TASK_MAX_CREDITS, BULK_TASK_INSUFFICIENT_CREDITS and LIST_TASK_ENGINE_ERROR
    • The MCP server renames list_jobs, get_job_results, get_job_rows and cancel_job to list_tasks, get_task_results, get_task_rows and cancel_task, bulk_enrich_persons takes taskName, and check_task replaces get_job: it follows a Bulk task (kind: bulk) as well as a live fetch or a contact lookup (kind: lookup)
    • The TypeScript and Python SDKs 0.4.0 send taskName (task_name in Python) and return data.task for Bulk and lists.importSignals, the Bulk types are now BulkTask, TaskResultLine and BulkTaskRows, and in Python the first argument of the client.bulk methods and of tasks.wait is task_id (previously job_id)
    • The CLI 0.8.0 calls the new routes, rc bulk create takes --task-name (previously --name), its error codes say task_ in place of job_, and rc task also reads a Bulk task (see CLI)

    ⚠️ Deprecated

    • In the TypeScript and Python SDKs, client.jobs, Job, JobFailedError and JobTimeoutError are deprecated aliases of client.tasks, Task, TaskFailedError and TaskTimeoutError, so existing code keeps working
  6. v2.45.2

    Bulk results as they come and as NDJSON, and rc bulk in the CLI

    You can now read the results of a Bulk job while it runs or download them as NDJSON, the job.updated webhook is retried for about 3 hours, and the CLI runs Bulk jobs and fills your Lists.

    ✨ Features

    • You can now read the results of a Bulk job a page at a time from the start of the run with GET /v2/jobs/{jobId}/rows, up to 100 items per page, for free (see Read job results by page)
    • GET /v2/jobs/{jobId}/results?format=ndjson returns one JSON object per item, with the person in the shape of /v2/enrich/persons, and the CSV stays the default (a job whose results were stored before this release keeps its CSV only)
    • The MCP server adds get_job_rows, which reads a page of results into the conversation even while the job runs, and get_job_results takes format for an NDJSON file
    • The TypeScript and Python SDKs add bulk.rows, bulk.iterateRows (bulk.iter_rows in Python) and NDJSON results in 0.3.0
    • The CLI, now at 0.7.0, adds rc bulk to start, follow, cancel and download Bulk jobs from a CSV, NDJSON or JSON file, and rc lists read, rc lists add and rc lists export for your Lists (see CLI)

    🚀 Improvements

    • A job.updated delivery that fails or times out is now tried again for about 3 hours, so the same event can arrive twice: key it on jobId and status

    🐛 Bug Fixes

    • In the results of a cancelled job, an item that was still running now reads running until it ends, within the hour, instead of an incomplete partial row
  7. v2.45.1

    Faster webhook delivery and a performance page

    Asynchronous jobs now start as soon as they are queued, so their results arrive about 10 seconds sooner, and a new page publishes the response times, webhook delivery and availability of the API.

    ✨ Features

    • You can now check the response times of each endpoint at P50, P90 and P99, the webhook delivery time and the share of requests the API served, measured on production traffic, on API performance and availability

    🚀 Improvements

    • Since October 27, an asynchronous job starts as soon as it is queued, and the median time from the request to its webhook fell from about 13 s to under 4 s
    • The faster start applies to /v2/fetch/persons/live, /v2/fetch/companies/live, /v2/contact/email and /v2/contact/phone, whether you receive the result by webhook or poll GET /v2/webhooks/{webhookId}
  8. v2.45.0

    Lists, Bulk and the job history, in the API, the MCP server and the SDKs

    You can now keep people and companies in Lists and buy their emails, phone numbers and profiles, enrich up to 25,000 people in one request and download the results as CSV, and follow every job of your workspace, from the REST API, the MCP server or the SDKs.

    ✨ Features

    • You can now create, list, read and delete the Lists of your workspace with GET /v2/lists, POST /v2/lists, GET /v2/lists/{listId} and DELETE /v2/lists/{listId}, all free
    • POST /v2/lists accepts an Idempotency-Key header, so a retried creation returns the List it created instead of a second one
    • You can add up to 1,000 people or companies in one call with POST /v2/lists/{listId}/members and get one result per Signal, in request order: a new member costs 1 credit from a Social profile URL and 2 from an email or a name, a member already on the List is free
    • You can import up to 25,000 Signals or a public CSV file or Google Sheet with POST /v2/lists/{listId}/intake/signals, and buy the profile, company, contact_email and contact_phone families for each member under a maxCredits ceiling
    • An import paused at its maxCredits or by a low balance keeps its work and resumes when you send it again with the same operationId, without charging anything twice
    • You can follow an import for free with GET /v2/lists/{listId}/operations/{operationId}/progress, with ?wait=20 long polling and per-line results that carry your externalId, or be told by the webhookUrl of the import each time it stops moving
    • You can cancel an import with POST /v2/lists/{listId}/operations/{operationId}/cancel from any key of the workspace: lines that had not started are never charged
    • You can read the members of a List page by page, find them by your own externalId, remove them, and export the List as CSV with GET /v2/lists/{listId}/export, one export at a time per workspace
    • API keys have a new lists permission, and a List call that spends credits also needs the permission of what it buys
    • You can enrich up to 25,000 people in one request with POST /v2/enrich/persons/bulk, with their profile, an email or a phone number, each item priced like its unit call and refunded when nothing is found (see Bulk)
    • POST /v2/enrich/persons/bulk accepts an Idempotency-Key header, so a retried request returns the job it created instead of charging a second one
    • You can follow a job for free with GET /v2/jobs/{jobId} and ?wait=20 long polling, or be told on its webhookUrl with a job.updated event once it settles
    • You can download the results of a job as CSV with GET /v2/jobs/{jobId}/results, one row per item in the order sent, with your externalId and what each item cost
    • You can cancel a job an API key or an AI assistant started with POST /v2/jobs/{jobId}/cancel: items that had not started are never charged
    • You can read every job of your workspace with GET /v2/jobs, List imports and Search exports started in the app included, filtered by type, status or List
    • The MCP server adds ten List tools, from create_list and import_to_list to export_list, and five Bulk tools: bulk_enrich_persons, list_jobs, get_job, get_job_results and cancel_job
    • The TypeScript and Python SDKs add client.lists and client.bulk in 0.2.0, with wait helpers that follow an import or a job until it stops moving
    • The MCP server adds four goal tools, create_goal, list_goals, get_goal and update_goal: a goal carries a task on your Lists across conversations, with criteria checked on the Lists and a credit budget that billed tools spend within when they receive its goalId (see Goals)

    🚀 Improvements

    • A JSON body over 25 MB is refused with 413 REQUEST_BODY_TOO_LARGE before it is read, at no charge
    • MCP tools now check the permissions of the API key the client connects with, as the REST API does
    • The CLI, now at 0.6.0, adds rc search organizations in place of rc search companies, --type on rc contact email to find a personal email, and --id on rc contact email and rc contact phone
    • rc search people and rc search organizations take --area with a zone id, rc search people takes --connections-count, and rc usage shows the credit limits of the API key
  9. v2.44.0

    Verify Email, search autocomplete for your product, a free trial without a time limit, and one answer for a blocked person

    You can now check whether an address can receive mail before you send to it, and build autocomplete on skills into the search of your own product. People Search can target one school exactly, and the free trial no longer ends after 7 days: it runs until its credits are spent. A person or company blocked under GDPR or CCPA now answers 451 on profile and company lookups, at no charge.

    ✨ Features

    • POST /v2/contact/email/verify checks an email address and returns valid, risky, invalid or unknown, for 1 credit per verdict. It is asynchronous like Find Email, by polling or webhook, and a check that fails is refunded
    • POST /v2/search/people/auto-suggestion returns the skills that start with what someone has typed, most common first, for a search box in your own product. It costs 0 credits and has its own rate limit, 1,000 calls per minute per workspace or 500 on the free trial, so typing never uses up your requests per minute.
    • People Search and its free count accept schoolId and schoolLinkedinId to target one school exactly
    • The TypeScript and Python SDKs add contact.verifyEmail and contact.verify_email in 0.1.7, then search.peopleAutoSuggestion and search.people_auto_suggestion in 0.1.8, which also accepts the new filters

    🚀 Improvements

    • A 422 on a field with a closed list now names the allowed values, for example "emailType" must be one of "professional", "personal", "any"
    • A new free trial has no end date: its 500 credits stay usable at 50 requests per minute until they are spent (previously 7 days)
    • A trial with no credit left gets 402 NO_CREDITS on /v2/search/people/count and /v2/search/organizations/count, as it already did on the free check endpoints
    • Within 2 hours of its last credit, a trial workspace pauses and every call returns 403 WORKSPACE_DISABLED, until a purchase of credits or a plan resumes it
    • A person or company blocked under a GDPR or CCPA data subject request answers 451 DATA_SUBJECT_BLOCKED at 0 credit on profile and company lookups, whichever input names them (profile URL, email, internal id, domain or name), and on Find Email and Find Phone by profile URL or internal id
    • In a Bulk Import, a blocked row ends as data_subject_blocked and is not charged
    • An async job that names a blocked person or company ends in error with the code data-subject-blocked, with no result, and its credits are refunded
  10. v2.43.0

    Enrich a company from its domain over MCP

    The MCP tool enrich_company now finds a company from its domain, as /v2/enrich/companies does.

    🚀 Improvements

    • The MCP tool enrich_company accepts a domain and returns the full company profile for 2 credits when a company is found and 0 when none is. A Social url still costs 1 credit and is used when both are sent. See Connect AI agents
  11. v2.42.0

    Search people and companies by region, market or economic group

    People search and Company search take a whole zone in one filter: a main region such as EMEA, a UN region or subregion, a market cluster such as DACH or the Nordics, or an institutional group such as the European Union, from a catalogue of 62 zones.

    ✨ Features

    • /v2/search/people, /v2/search/organizations and their /count endpoints accept location.area, one zone id or an array, such as "MR:EMEA", ["MC:DACH", "MC:NORD"] or "IG:EU". People search matches the person's country, Company search the country of the headquarters. The 62 zones and their member countries are listed in Area values
    • Zones add up with each other and with location.countryCode and location.country, while location.region and location.city narrow the result: { "area": "MC:DACH", "countryCode": "FR" } covers Germany, Austria, Switzerland and France. A zone only adds countries, there is no exclusion
    • The available-fields endpoints list the zone ids in the allowedValues of location.area, and each zone with its member countries in areaCatalog
    • The prompt endpoints turn a region named in the prompt, such as "in the Nordics" or "across EMEA", into its zone id
    • The MCP tools search_persons, count_persons, search_companies and count_companies accept location.area
    • /v2/search/people and /v2/search/people/count accept connectionsCount, a connection count range where one bound is enough: { "min": 501 } keeps profiles with more than 500 connections. The MCP tools search_persons and count_persons and the prompt endpoint take it too
    • Both SDKs, from version 0.1.6, accept location.area on search.people, search.organizations and their counts, and connectionsCount on People search
    • Both SDKs, from version 0.1.6, publish the values of the fixed lists as constants: Industry, CountryCode, JobFunction, JobSubFunction, Seniority, ContractType, Area and EmailType, such as Area.MC_DACH or Industry.SOFTWARE_DEVELOPMENT. A filter still accepts any string. See the JavaScript SDK and Python SDK pages

    🚀 Improvements

    • A zone acronym or name sent in location.area, such as "DACH", returns HTTP 422 with INVALID_FILTER_VALUE and the full ids to use in details.suggestions, and costs 0 credits. MENAT, MENAP and SAARC also return the countries to add to their zone in details.countryCodes
  12. v2.41.0

    Official SDKs, person ids and safe retries on Find Email and Find Phone

    Reverse Contact now has official SDKs for TypeScript and JavaScript and for Python. Find Email and Find Phone accept an Idempotency-Key header, so retrying a request never starts a second lookup, and take a Reverse Contact person id, so a person you already found with Search, Fetch or Enrich goes straight to the contact lookup.

    ✨ Features

    • @reverse-contact/sdk is the official TypeScript and JavaScript SDK, on npm, for Node.js 22 and later, Bun, Deno and edge runtimes, with no runtime dependency: see TypeScript and JavaScript SDK
    • reversecontact is the official Python SDK, on PyPI, for Python 3.11 and later, with a synchronous client and an asynchronous one: see Python SDK
    • Both SDKs retry per-minute rate limits and server errors on their own, never retry daily limits or API key credit limits, and send an Idempotency-Key on Find Email and Find Phone so a retry never starts a second lookup
    • Both SDKs wait for live fetch, Find Email and Find Phone jobs with wait(), return null or None when an enrich or fetch lookup finds nothing, and page through search results within a maxResults or max_results budget
    • Every endpoint page now shows its call with each SDK, in the TypeScript SDK and Python SDK tabs of its code examples
    • POST /v2/contact/email and POST /v2/contact/phone accept an optional Idempotency-Key header. For 24 hours, a request sent again with the same key and the same body returns the first response, with the same webhookId and the header Idempotent-Replayed: true, without starting another lookup or debiting credits
    • Three error codes come with it: 409 IDEMPOTENCY_KEY_IN_USE while the first request with the key is still being processed (retry after Retry-After), 422 IDEMPOTENCY_KEY_REUSED when the key was already used with a different body, and 400 INVALID_IDEMPOTENCY_KEY for a key that is not 1 to 255 printable ASCII characters
    • POST /v2/contact/email and POST /v2/contact/phone accept id, the prs_… id of any person returned by Search, Fetch or Enrich, in place of url or a name with companyDomain. If both url and id are supplied, id takes precedence. Prices do not change
    • An id that matches no person returns 422 VALIDATION_ERROR before any lookup starts, and costs 0 credits
    • Both SDKs, from version 0.1.5, accept id on contact.findEmail and contact.findPhone (contact.find_email and contact.find_phone in Python), and the MCP tools find_email and find_phone accept it too

    🚀 Improvements

    • Find Email and Find Phone return 451 DATA_SUBJECT_BLOCKED for a person removed per a GDPR/CCPA request, whether the person is given by url or by id, as the other person endpoints do. The request costs 0 credits
  13. v2.40.0

    Credit limits per API key

    Each API key can now cap the credits it spends, per day, per month or in total.

    ✨ Features

    • API keys accept three optional credit limits: per day (UTC), per month (UTC) and in total, set from the API Keys page when you create or edit a key. A total limit counts from the moment you set it and starts over only when you reset it
    • A call that would take a key over one of its credit limits is refused before it runs and charges nothing, with the code KEY_CREDIT_LIMIT_REACHED: 429 with Retry-After for the daily and monthly limits, 403 for the total limit
    • Every response made with an API key carries quotas.key.creditLimits, with limit, used, left and nextReset for each limit the key has, this call included
    • The limits apply on every channel that uses the key: the API, the MCP server and the CLI

    🚀 Improvements

    • The API Keys table shows the credit limit each key is closest to reaching, and the edit window shows what the key spent today, this month and against its total
    • The per-key "Daily limit" is now called "Daily request limit", since it counts requests, not credits
  14. v2.39.0

    Contract types and schools in People Search

    People Search can now target how people are employed and where they studied.

    ✨ Features

    • People Search and its free count accept currentContractType and contractType, on the current position or on any position, with eight values such as freelance, internship or part-time
    • People Search and its free count accept excludeCurrentContractType and excludeContractType to refine a search, next to a positive filter
    • People Search and its free count accept schoolName: every word must appear in the name of one school the person attended, so "HEC Paris" never matches words spread across two schools
    • Agentic Search now reads contract types and schools from a prompt, such as freelance designers who studied at ESSEC, instead of reporting them as not applied
    • The People Search MCP tools search_persons and count_persons accept the same five filters

    🐛 Bug Fixes

    • An any Email Finder lookup that finds both a professional and a personal email now reports creditsConsumed: 8, the credits it debited, on its last poll of GET /v2/webhooks/{webhookId} and in the MCP check_task tool. It reported 5, while your balance was debited correctly
  15. v2.38.0

    Faster async results with long polling

    Polling an async job can now wait on the server, so you get the result as soon as the job ends instead of on your next poll.

    ✨ Features

    • GET /v2/webhooks/{webhookId} accepts a wait parameter (0 to 20 seconds): the poll answers as soon as the job ends, or after wait seconds with the job still running, and stays free and not rate limited

    🚀 Improvements

    • Async responses that return a webhookId, and every poll of a job still running, now carry a Retry-After header with the seconds to wait before the next poll
  16. v2.37.1

    Company search maintenance starts November 9

    /v2/search/companies goes into maintenance on Monday, November 9, 2026: move your company searches to /v2/search/organizations before that date.

    ⚠️ Deprecated

    • /v2/search/companies and /v2/search/companies/available-fields go into maintenance on November 9, 2026, for a few weeks: from that date they return 410 with ENDPOINT_GONE and consume no credits, so call /v2/search/organizations and /v2/search/organizations/available-fields instead
    • After the maintenance, /v2/search/companies comes back in a format compatible with /v2/search/organizations, so an integration moved to /v2/search/organizations needs no second migration
  17. v2.37.0

    Every endpoint on the free trial, with 500 credits

    The free trial now opens every endpoint, live endpoints included, and a new trial starts with 500 credits.

    🚀 Improvements

    • Every endpoint is available during the free trial, including /v2/fetch/persons/live and /v2/fetch/companies/live, which no longer return 403 TRIAL_ENDPOINT_RESTRICTED
    • The MCP live tools enrich_person_live and enrich_company_live work on a trial workspace and no longer return TRIAL_RESTRICTED
    • Bulk imports accept Live enrichment during the free trial, the dashboard opens its live pages to trial workspaces, and rc fetch ... --live works on a trial
    • A new free trial starts with 500 credits (previously 100). Trials already running keep their balance, and a trial still lasts 7 days at 50 requests per minute
  18. v2.36.1

    Company id on every person position

    Every person endpoint now returns the Reverse Contact id of the company on each position.

    🐛 Bug Fixes

    • /v2/fetch/persons and /v2/enrich/persons now return companyId on currentPosition and on every experience item, as People search already did, so you can pass it as id to /v2/fetch/companies
  19. v2.36.0

    Personal emails in the Email Finder

    The Email Finder now finds personal emails too: choose professional, personal or either, from the API, the dashboard, Bulk imports, Ask and the MCP server.

    ✨ Features

    • POST /v2/contact/email accepts emailType: professional, personal, or any to accept either kind. Without it the lookup still finds professional emails only, so existing integrations do not change
    • Every found email carries its type (professional or personal)
    • The price depends on the kind of email found, not on emailType: 3 credits for a professional email, 5 for a personal one, so an any lookup costs 3, 5 or 8, the 8 only when it finds both. Each kind is billed once, however many addresses come back, and a lookup that finds nothing stays free
    • A professional lookup (the default) still debits 3 credits when the job is created. personal and any debit 5 upfront, and an any lookup is adjusted when the result arrives: 2 credits back when only a professional email is found, 3 more when both kinds are found. Before accepting the request, the API checks that your balance could pay for the most expensive result (3 credits for a professional lookup, 5 for personal, 8 for any in case it finds both), and this check does not set the price
    • The Find Email page and Bulk imports that find emails let you pick the email type
    • Ask and the MCP find_email tool look for a personal email when you ask for one
  20. v2.35.1

    People Search drops the area filter

    People Search no longer accepts the location.area filter.

    💥 Breaking Changes

    • /v2/search/people, /v2/search/people/count and the MCP search_persons and count_persons tools no longer accept location.area: a request that sends it answers 422 with UNKNOWN_FIELD and costs 0 credits, so use location.city for a metropolitan area or location.region for a state or province
  21. v2.35.0

    Shape the roadmap, sharper Agentic People Search and CLI 0.3.0

    Paying workspaces can now shape the roadmap from the dashboard, Agentic Search finds people for far more prompts, and the CLI adds the phone finder and cursor pagination.

    ✨ Features

    • Shape the roadmap replaces the feature request board in the dashboard (under Help) for paying workspaces: answer the open product surveys, one answer per workspace, and see who answered and when
    • CLI 0.3.0 adds rc contact phone, which finds a person's phone numbers and costs 20 credits only when a number is found
    • rc search people paginates with cursors: pass --cursor with the nextCursor of the previous response and set the page size with --limit (up to 100)

    🚀 Improvements

    • Agentic Search for people now reads job titles written in the plural or in another language, places in their local spelling and company domains such as stripe.com as filters that match profiles, so far fewer prompts come back empty
    • A short prompt such as "recruiters" now runs a search instead of reading no filter, and a list of topics or a vague sector no longer narrows the search beyond what you asked
    • In the dashboard, a People description that yields no filter now explains what to add instead of leaving the field silent
    • Bulk CSV imports and search exports now apply your plan like single calls do: covered rows and exports cost 0 credits, and a covered job starts even with an empty credit balance
    • Bulk CSV import rows now count against your per-minute rate limit and use at most 80% of it, so your direct API calls keep the rest
    • The retired Resolve pages are removed from the dashboard, along with their rows in the API pricing table
    • The API pricing table labels Enrich Profile and Enrich Company as Hybrid (previously Live sourcing)

    🐛 Bug Fixes

    • A prompt made only of refinements, such as "open to work" or a follower count, is now refused with NO_FILTERS_INTERPRETED before the search and costs 0 credits, instead of failing at the search step
    • rc version now tells you when a newer CLI release exists (the check never reported one before)
    • The OpenAPI spec at https://api.reversecontact.com/openapi.json no longer lists the retired Resolve endpoints and states People Search at 1 credit per profile delivered
  22. v2.34.0

    Company Search with full profiles, free counts and Agentic Search

    Company Search returns full company profiles, paginates with cursors, filters on specialities, founding year and followers, can be sized for free before you spend credits, and now reads its filters from a plain-language prompt too.

    ✨ Features

    • Company Search is available at POST /v2/search/organizations and returns full company profiles directly, the same shape as the Company profile endpoint, with cursor pagination and perPage from 1 to 100 (default 25)
    • Company Search costs 1 credit per company delivered, and a page with no company or a rejected filter costs 0 credits
    • Company Search accepts three new filters: specialities (one value or a list, OR-matched), foundedYear and followersCount (whole number ranges), next to companyName, companyDomain, industry, employeeCountRange, location and maxDataAgeDate
    • POST /v2/search/organizations/count returns the total number of matching companies for the same filters, for free
    • GET /v2/search/organizations/available-fields lists every accepted company filter with its accepted values, for free
    • Agentic Search now covers companies too: POST /v2/search/organizations/prompt takes a prompt in any language, such as "Fintech scale-ups headquartered in Berlin, 200 to 500 people", and returns the matching companies as a Company Search page, with the filters applied and a warning for anything no filter covers, such as funding, in data.interpretation
    • Agentic Company Search costs 1 credit per page for the search plus 1 credit per company delivered, like Agentic Search for people, and a prompt refused before the search costs 0 credits

    ⚠️ Deprecated

    • /v2/search/companies and /v2/search/companies/available-fields are deprecated and succeeded by /v2/search/organizations and /v2/search/organizations/available-fields: they are still served and no retirement date is announced yet, but they have left the endpoint catalog
    • The migration guide covers every change, from the renamed result fields to the move from pages to cursors
  23. v2.33.0

    Resolve and persons search retired

    The previous persons search and Resolve endpoints are retired and now return 410 Gone: People search and Enrich replace them with the same inputs and costs.

    💥 Breaking Changes

    • /v2/search/persons and /v2/search/persons/available-fields now return 410 Gone (use /v2/search/people and /v2/search/people/available-fields)
    • /v2/resolve/persons/email and /v2/resolve/persons/name now return 410 Gone (use /v2/enrich/persons, which answers synchronously with the same input and cost)
    • /v2/resolve/companies/live now returns 410 Gone (use /v2/enrich/companies, which answers synchronously with the same input and cost)
    • The 410 response carries the error code ENDPOINT_GONE, with error.details.alternative naming the endpoint to call instead and error.details.docs linking the endpoint transition guide
    • Calls to a retired endpoint are not processed and consume no credits: move your persons searches to People search (/v2/search/people) and your Resolve calls to Enrich (/v2/enrich/persons or /v2/enrich/companies)
  24. v2.32.0

    Skills, languages, keywords and job changes in People Search

    People Search can now target what people know, the languages they speak, any word on their profile and who just started a new role.

    ✨ Features

    • People Search and its free count accept skills: one skill name or a list, matched as whole names, so "Java" does not return people who only list "JavaScript"
    • People Search and its free count accept languages and excludeLanguages with language names in English, which also match the spellings profiles use (French finds Français), and about one profile in six lists its languages
    • People Search and its free count accept keywords: every word must appear somewhere in the profile, from the headline and summary to past job titles, schools and location
    • People Search and its free count accept changedJobsWithinDays with 30, 60, 90 or 120 to find people who started their current position within that many days
    • Agentic Search now reads skills, spoken languages, recent job changes and free keywords from a prompt instead of reporting them as not applied
    • The People Search MCP tools search_persons and count_persons accept the same four filters
  25. v2.31.0

    People Lookalike, email and phone finding in MCP

    Your AI agent can now find people like a profile and look up their professional email addresses and phone numbers through the MCP server.

    ✨ Features

    • The MCP tool search_persons_lookalike finds people who look like one seed profile, given by its Social profile URL, and returns full profiles shaped like search_persons with cursor pagination up to 100 per request
    • search_persons_lookalike costs 1 credit per profile delivered, and an empty page or a seed that is not in the search index yet costs 0 credits
    • The MCP tools find_email and find_phone find a person's professional email addresses and phone numbers from a Social profile URL, or from a name with companyDomain, and return a taskId to poll with check_task
    • find_email costs 3 credits and find_phone 20 credits, debited when the job is created and refunded when nothing is found
    • A workspace without phone finding access gets CONTACT_PHONE_FINDER_DISABLED from find_phone and is not charged
    • The MCP setup guide lists the 16 tools with their credits, and documents the new error codes
  26. v2.30.0

    Agentic Search API

    Describe the people you are looking for in one sentence and get their full profiles back, with no filters to build.

    ✨ Features

    • Agentic Search is available at POST /v2/search/people/prompt: send a plain-language prompt in any language, up to 1000 tokens, and get back full profiles as People Search returns them
    • Every response carries data.interpretation, with the People Search filters read from the prompt and warnings for anything the search could not apply, such as skills or schools
    • Agentic Search uses cursor pagination with perPage from 1 to 100 (default 25): send the same prompt with metadata.nextCursor to get the next page, which runs on the same filters
    • Agentic Search costs 1 credit per page for the search itself, plus 1 credit per profile delivered, and a prompt refused before the search, for example one no filter can be read from, costs 0 credits
    • The Agentic Search page documents the request, the filters a prompt can express, the response and every error
  27. v2.29.0

    People Lookalike

    Start from one profile you know and find the people who look like it, from the API or from the dashboard.

    ✨ Features

    • People Lookalike is available at POST /v2/search/people/lookalike: send one seed profile, as a Social URL (url) or a Reverse Contact person id (id), and get back the full profiles of people with a similar current function, sub-function, seniority and location
    • People Lookalike uses cursor pagination with perPage from 1 to 100 (default 25), and never returns the seed itself
    • People Lookalike costs 1 credit per profile delivered, and a seed that is not in the search index yet returns 404 for 0 credits
    • The People lookalike page documents the request, the response and every error

    🚀 Improvements

    • The Lookalike page of the dashboard accepts a Reverse Contact person id (prs_...) as the seed, as Fetch does, and keeps it in the page address so the link can be shared
  28. v2.28.0

    Retirement of Resolve and persons search on October 13, 2026

    The previous persons search and Resolve endpoints are retired on October 13, 2026: migrate to People search and Enrich before that date.

    ⚠️ Deprecated

    • /v2/search/persons and /v2/search/persons/available-fields are retired on October 13, 2026 (use /v2/search/people and /v2/search/people/available-fields)
    • /v2/resolve/persons/email and /v2/resolve/persons/name are retired on October 13, 2026 (use /v2/enrich/persons, which answers synchronously with the same input and cost)
    • /v2/resolve/companies/live is retired on October 13, 2026 (use /v2/enrich/companies, which answers synchronously with the same input and cost)
    • The endpoint transition guide and the new Migrating from Resolve to Enrich guide cover every change step by step
  29. v2.27.0

    Full profile by default on Enrich endpoints

    Enrich Person and Enrich Company now return the full profile in every response, at the same 2 credits.

    🚀 Improvements

    • /v2/enrich/persons and /v2/enrich/companies now return the full profile by default (same fields as /v2/fetch/persons and /v2/fetch/companies) for 2 credits when a match is found, with no extra credit
    • The alternative candidates in data.alternativePersons come back as full profiles too, at no extra credit
    • The response stays compatible: every field of the previous light profile keeps its name, place and type, and the full profile only adds fields
    • A new data.isFullProfile field is true when the full profile is returned, and false in the rare case where only the light profile could be retrieved in time (you can then complete it with /v2/fetch/persons or /v2/fetch/companies)
    • Not-found responses (404) stay free

    ⚠️ Deprecated

    • The fullProfile request field is deprecated and ignored: it is still accepted so existing requests keep working, and you can remove it
    • The --full-profile flag of rc enrich person and rc enrich company and the fullProfile parameter of the MCP enrich_person tool are still accepted but ignored
  30. v2.26.0

    New People Search with full profiles and free counts

    People search returns full profiles, paginates with cursors, and can be sized for free before you spend credits.

    ✨ Features

    • People Search is available at POST /v2/search/people and returns full profiles directly, with cursor pagination and perPage from 1 to 100 (default 25)
    • People Search accepts position taxonomy filters: currentFunction across 32 functions, currentSubFunction across 386 sub-functions and currentSeniority across 10 seniorities, plus past-experience requirements with a minimum tenure
    • People Search also matches current and past positions through the positionTitle, companyName, companyId, companyLinkedinId, companyPublicId and companyIndustry filters, and accepts an email filter
    • POST /v2/search/people/count returns the total number of matching people for the same filters, for free
    • GET /v2/search/people/available-fields lists every accepted people filter with its accepted values, for free

    🚀 Improvements

    • People Search costs 1 credit per profile delivered, and a page with no result or a rejected filter costs 0 credits
    • The previous people search path /v2/search/persons and its available-fields route keep working during the migration but have left the endpoint catalog, and the migration guide covers every change
  31. v2.25.0

    Find Email is generally available

    Find Email leaves beta and is now generally available on every plan.

    🚀 Improvements

    • Find Email is now generally available and is no longer marked as beta
  32. v2.24.0

    Search pagination capped at 50 pages

    Search pagination now reaches the first 50 pages instead of 500, while perPage and search exports stay as they are.

    💥 Breaking Changes

    • page on /v2/search/persons and /v2/search/companies now accepts 1 to 50 (previously 1 to 500), which reaches the first 5,000 matching records at the default perPage of 100
    • A page above 50 returns a 422 PAGINATION_LIMIT_EXCEEDED whose details carries field set to page and max set to 50, the same shape already used for perPage
    • pageNumber and total still describe the whole matching set, so a search can report more than 50 pages while only the first 50 are retrievable, and narrower filters are the way to reach the rest
    • perPage is unchanged on both endpoints and still accepts up to 100 results per page, and search exports are unaffected
  33. v2.23.0

    August round-up: Contact Email Finder, Claude directory, Ask, search guards

    ✨ Features

    • Contact Email Finder is available in public beta on every plan, including trials, with an API key scoped to contact
    • Contact Email Finder accepts a profile URL, fullName plus companyDomain, or firstName and lastName plus companyDomain, with an optional webhookUrl
    • Contact Email Finder returns results asynchronously through GET /v2/webhooks/:id or a callback as [{value, type}], where type is professional or personal
    • Contact Email Finder returns only verified addresses, and unsuccessful lookups return email-not-found instead of an unverified candidate
    • Contact Email Finder costs 3 credits only when an email is found, and 0 credits for email-not-found, person-not-found, invalid-request, and fetch-data-error
    • A Contact Email Finder delivery failure after a match remains billed, and the result stays available through GET /v2/webhooks/:id
    • Contact Email Finder is available in Ask as contact_email_find and in the CLI as rc contact email, but not in MCP
    • You can add Reverse Contact to Claude from https://claude.ai/directory/reverse-contact in one click, without creating a custom connector
    • Claude can enrich people and companies with cached or live data, search people and companies, check a task, and view usage, at the same credit pricing as the API
    • Live mode on /v2/fetch/persons/live and /v2/fetch/companies/live accepts internal prs_... and com_... ids as alternatives to profile URLs
    • An id that cannot be resolved returns 422 rather than not-found, and the live fetch cost remains 2 credits

    🚀 Improvements

    • Ask keeps watching live enrichment jobs that outlast a conversation turn, polling every 3 seconds and then every 6 seconds after 60 seconds for up to 10 minutes
    • Ask inserts the result into the thread when the job completes, or shows a Resume button when the monitoring budget ends
    • Stopping Ask monitoring never cancels the job, and the job is billed when it completes
    • Search going past the last page now returns a clear error instead of an empty billed page
    • Unknown search filter keys and invalid filter types return 422 with 0 credits charged
    • Search documents employee count ranges for people and company filters, and list filters accept at most 20 values per field
    • Error codes are documented end to end in the MCP and CLI documentation, with code documentation added earlier in August
    • Requests that exceed their deadline return structured JSON with 504 status, cost 0 credits, and are safe to retry; 503 remains reserved for real service unavailability

    💥 Breaking Changes

    • Search pagination is now capped at 500 pages, down from 1,000; requests beyond that return a 422 validation error and are not charged
  34. v2.22.0

    One-click ChatGPT MCP listing

    Reverse Contact is now listed in the ChatGPT plugin directory, so you can connect MCP without creating a custom connector.

    ✨ Features

    • You can now add Reverse Contact to ChatGPT in one click from the ChatGPT plugin directory, with no API key and no custom connector
  35. v2.21.0

    Ask exports and stricter input validation

    Ask results can now leave the chat as a file, and V2 endpoints reject unusable input up front instead of answering with an empty match.

    ✨ Features

    • You can download any Ask result as CSV or JSON from the result card, at no credit cost
    • In Ask, a plain-language request such as "export these results to CSV" prepares the file from what is already on screen
    • Export everything a conversation returned in a single file, with duplicate people and companies removed
    • Exports only ever contain results already returned, so a search exports the page it answered with and widening it still means running a new search

    🚀 Improvements

    • V2 endpoints now trim surrounding whitespace and strip invisible characters from every text field, so a value copied from a spreadsheet no longer fails to match for an invisible reason
    • Email addresses are lowercased automatically on V2 endpoints
    • API keys are trimmed before validation, so a key pasted with a trailing space no longer returns an authentication error

    🐛 Bug Fixes

    • Changing your account email now states clearly that both the current and the new address have to confirm before the change takes effect

    💥 Breaking Changes

    • V2 endpoints now reject unusable identity values with a 422 instead of returning no match: placeholders such as [not provided] or n/a (INVALID_PLACEHOLDER_VALUE), spreadsheet error tokens such as #N/A or #REF! (SPREADSHEET_ERROR_VALUE), an email address in a name field (EMAIL_IN_NAME_FIELD), and HTML tags (INVALID_CHARACTERS), all at 0 credits
    • People endpoints return a 422 MISSING_PERSON_IDENTIFIER when only a company domain or company name is provided, and point you to /v2/enrich/companies instead
    • A webhookUrl that is not publicly reachable, such as localhost or a private IP address, is now rejected with a 400 WEBHOOK_URL_UNREACHABLE on every V2 endpoint that accepts one (previously accepted)
  36. v2.20.0

    Faster live enrichment for CSV imports

    Live enrichment on CSV imports now runs many more profiles in parallel, and large jobs take minutes instead of hours.

    🚀 Improvements

    • Live enrichment on CSV imports now processes up to 150 profiles at once (previously 10), so an 8,000 row import finishes in minutes instead of hours
    • Rows are picked up faster once a live scrape completes, so there is no idle time between profiles
    • Per-profile scrape time and credit cost are unchanged: only the number of profiles running in parallel increased
  37. v2.19.0

    Live profile rescrapes in Ask

    Ask can now live-rescrape people and company profiles when you ask for fresh data, with a free follow-up when a scrape is still running.

    ✨ Features

    • Ask can live-rescrape a person or company Social profile when you explicitly ask for fresh, live, or up-to-date data, instead of only using the cached profile
    • Live rescrapes run asynchronously: Ask waits for the result when it can, and if the job is still running you get a task id and can continue with a free check_task follow-up
    • Live jobs target under 15 seconds but can take several minutes depending on conditions; credits are billed when the job completes, and check_task never costs credits
  38. v2.18.0

    Live enrichment for bulk CSV Social URL rows

    Bulk imports can now fresh-scrape Social URL rows when you opt in to Live enrichment on the Review step.

    ✨ Features

    • When the mapping of a bulk CSV import includes a Social URL column, you can enable Live enrichment on the Review step to fresh-scrape those URL rows instead of using cached match data
    • Live URL rows are async only: each scrape targets under 15 seconds, and a full job can take several minutes depending on size and conditions
    • Live URL rows cost 2 credits only when the row succeeds; not-found rows cost 0 credits, and reserved surplus is settled when the job completes
    • Email and name rows stay on the standard path and pricing, even when Live enrichment is enabled on the job
    • Live jobs show a Live badge in import history and on the job detail page
    • Live enrichment is not available on free trial workspaces
  39. v2.17.0

    CLI 0.2: fetch, check, contact, live jobs, and richer search

    The Reverse Contact CLI (rc 0.2.0) adds cheaper URL fetches, free existence checks, async live scrapes, contact tools, and more search filters to the terminal.

    ✨ Features

    • Fetch a person or company by Social URL or internal id with rc fetch person and rc fetch company (1 credit when cached, cheaper than enrich when you already have a URL)
    • Check whether a profile exists and how fresh it is for free with rc check person and rc check company (0 credits)
    • Trigger a live scrape with rc fetch … --live, poll the job with rc task <taskId>, or pass --wait when you want the CLI to block until it finishes
    • Find and verify contact emails from the CLI with rc contact email and rc contact verify
    • Discover server-side search filters with rc search fields people and rc search fields companies so agents stay aligned with the API schema
    • Install on macOS with Homebrew: brew install visum-run/tap/rc

    🚀 Improvements

    • rc enrich person and rc enrich company accept --full-profile for the full profile shape (+1 credit when a match is found)
    • People and company search expose more filters: multi-value country, city and industry, company ids, open-to-work, premium, verified, followers range, and max data age
    • rc auth status and rc usage surface rate limit and daily limit alongside credit balance
  40. v2.16.0

    Social URL column for bulk CSV enrichment

    Bulk CSV enrichment now accepts a Social profile URL as an input column and enriches straight from the URL at a lower cost per row.

    ✨ Features

    • Bulk CSV import now accepts a Social profile URL column, auto-detected from headers like linkedin_url, linkedin, profile_url or url, or mapped manually in the Map step
    • Sales Navigator links and messy URLs, including query strings, a missing scheme, or a bare slug, are normalized automatically before enrichment

    🚀 Improvements

    • A row enriched directly from a valid Social profile URL now costs 1 credit on success, instead of 2 for the standard email or name based match, and rows that fail or come back not found are never charged
    • A row that points to a company page is not a valid person profile, so it is skipped, marked invalid_input in the result CSV, and never charged
    • Duplicate detection across rows now recognizes the same profile across URL variants, and a matching Social URL takes priority over a matching email when both are present
  41. v2.15.0

    Full profile on Enrich endpoints

    Enrich Person and Enrich Company can now return the full profile in a single call, without a second call to the fetch endpoints.

    ✨ Features

    • You can now pass fullProfile: true on /v2/enrich/persons and /v2/enrich/companies to get the full profile (same completeness as the fetch endpoints) for 1 extra credit
    • The extra credit is only charged when a match is found and the full profile is delivered; not-found responses stay free, even with fullProfile
    • On Enrich Person, the full profile adds fields such as summary, photoUrl, experience, education, skills, languages and certifications, while alternativePersons keeps the light shape
    • Default behavior is unchanged: without the flag you get the light profile at the current cost
  42. v2.14.0

    Alternative candidates on Enrich Person

    The Enrich Person endpoint now surfaces the other candidates it considered, so you can see who else matched before settling on the top result.

    ✨ Features

    • Enrich Person now returns alternativePersons, the other candidates that matched your input, ranked best first
    • Alternative candidates are visible in the dashboard right below your enriched profile, with a link to each one's Social profile
    • alternativePersons is mostly populated when you enrich by email or by name and company, and empty when the match is unique, which is typical for a Social URL lookup
  43. v2.13.0

    Clearer delivery expectations for live endpoints

    This release clarifies how live endpoints are meant to be used: the docs and the live lookup form now spell out realistic delivery timing and the async-first design.

    🚀 Improvements

    • Live endpoints now state a clear delivery expectation across the docs and the lookup form: results target under 15 seconds but can take tens of minutes depending on conditions
    • Clarified that live endpoints are built to update databases and feed async pipelines, not to sit inside a synchronous real-time request where a user is waiting on the response
  44. v2.12.0

    Trial feedback rewards, bulk usage visibility and API key fixes

    This release rewards feedback after your trial ends, makes bulk CSV usage visible across your stats and logs, adds people search by current company, and fixes API key permissions on the Enrichment endpoints.

    ✨ Features

    • Share feedback after your trial ends and receive 50 bonus credits plus a 7-day trial reactivation on your first submission
    • Filter people searches by current employer using a company id (com_…) in the Company ID field, matching only the person's current company
    • A new Integrations section in the sidebar groups the MCP server, CLI and other ways to connect
    • Bulk CSV enrichments now appear as a dedicated Bulk CSV channel in usage analytics and as a matching source in your request logs

    🚀 Improvements

    • Webhook enrichments where the data was fetched but delivery to your webhookUrl failed are now returned as succeeded with a new deliveryStatus field, so you no longer discard a result you already paid for
    • Refunded credits from failed async enrichments are now reflected in your usage analytics and the Top Endpoints breakdown
    • Support replies now render Markdown formatting, both in-app and in email
    • You can set a password on your account even if you signed up with Google or a magic link, so email and password sign-in works as a backup
    • The Contact tool is merged into Enrich, so you can resolve a person from their email address on the Enrich page
    • The API playground now has default values, dynamic search filters and a mobile-friendly layout
    • Pasting a Social profile URL on the Enrich page now sends you to Fetch, which handles URL lookups

    🐛 Bug Fixes

    • Granular and resolve-only API keys can now call the /v2/enrich/persons and /v2/enrich/companies endpoints, which previously returned a 403 permissions error
    • Polling an async result at /v2/webhooks/:webhookId with a granular API key no longer returns a 403
    • The search filter panel no longer opens automatically when you load the search page
  45. v2.11.0

    Bulk enrichment, search exports to CSV and MCP sign-in

    This release introduces bulk enrichment from CSV files or pasted lists, CSV export for search results, and account sign-in (OAuth) for the MCP server.

    ✨ Features

    • On the new /bulk page, you can upload a CSV or paste a list of emails or names, map your columns, and enrich up to 25,000 people in one bulk enrichment job
    • Follow each bulk job live with row-by-row progress, a breakdown of enriched, not found and invalid rows, and a results CSV ready to download
    • When your credit balance runs low mid-job, bulk jobs enrich as many rows as your balance allows instead of failing the whole job
    • You receive an in-app notification and an email when a bulk job completes, with a direct link to the results
    • People and company searches now have an Export CSV button, and a new exports page keeps your past exports available for download
    • Claude clients (claude.ai, Claude Desktop, Claude Code, mobile) can now connect to the MCP server when you sign in with your account, through OAuth with a consent page and no API key to copy
    • Each MCP OAuth connection creates a dedicated API key named MCP - {app} in your workspace, which you can revoke anytime from the API Keys page

    🚀 Improvements

    • The CLI install page now detects your operating system and shows the matching one-line install command, with a new install script for Windows
    • The MCP setup page now includes one-line terminal commands to connect Claude Code and other CLI clients
    • Trial workspaces can now run bulk enrichments and search exports
    • /v2/fetch/persons responses now include the person's Reverse Contact id (prs_…), and ids are documented across fetch, enrich, resolve and search responses
    • API latency stats now include P50, P90 and P95 percentiles

    🐛 Bug Fixes

    • Retired v2 activity endpoints now return a proper 410 JSON error pointing to the sunset guide, instead of an HTML error page

    💥 Breaking Changes

    • Sign-in with GitHub is removed, and accounts that used it can sign in with Google or with email and password
  46. v2.10.0

    Fetch by Reverse Contact id and discover available fields

    This release lets you fetch profiles and companies by their Reverse Contact id, adds endpoints to discover every field you can request, and points Resolve users toward synchronous Enrichment.

    ✨ Features

    • You can now fetch a person or company by its Reverse Contact id, passing id (prs_… or com_…) to /v2/fetch/persons and /v2/fetch/companies instead of the Social URL
    • The dashboard status check now accepts a Reverse Contact id and runs the lookup automatically
    • The new available-fields endpoints, /v2/search/persons/available-fields and /v2/search/companies/available-fields, list every field you can request at no credit cost

    🚀 Improvements

    • Resolve pages now point you toward synchronous Enrichment, with a new "Resolve or Enrichment?" guide in the docs (Resolve stays fully supported)
    • The MCP enrich_person tool now accepts a work email in addition to a Social URL
    • The enrich profile form now guides you toward work emails for better match rates
    • The MCP and CLI setup page now includes troubleshooting sections
  47. v2.9.0

    Reverse Contact CLI and optional two-factor authentication

    This release ships the official Reverse Contact CLI, adds optional two-factor authentication to your account, and reduces the MCP server to 6 tools.

    ✨ Features

    • You can install the Reverse Contact CLI (the rc binary) with a single command from the new install page at /cli, with signed binaries for macOS, Linux and Windows
    • rc enrich person and rc enrich company resolve profiles and companies from your terminal, with the same inputs and pricing as the API
    • rc search people and rc search companies search from your terminal and return JSON output for scripting
    • rc auth login signs you in through a browser device flow you approve from your workspace, with no key to copy and paste, and rc auth status and rc auth logout manage the session
    • rc usage shows your credit consumption and rc version checks daily for new releases
    • Run npx skills add visum-run/skills to install the CLI agent skill, which lets coding agents drive Reverse Contact from the terminal
    • You can enable optional two-factor authentication with an authenticator app from your account security settings, and your sign-ins then require a 6-digit code (nothing changes if you don't enable it)

    🚀 Improvements

    • You can now create up to 25 active API keys per workspace (previously 10)
    • People search results now display each person's location
    • Every "persons" label across the dashboard and docs now reads "people", and endpoint URLs are unchanged
    • Expired trials now show a clear trial-expired message instead of a generic workspace-disabled error
    • Docs navigation now lists Search above Fetch to match how most integrations start

    💥 Breaking Changes

    • The MCP server now exposes 6 tools covering enrich, search and utilities, and the resolve, contact and activity tools are removed

    ⚠️ Deprecated

    • The Contact API family is now marked as deprecated in the docs and navigation
  48. v2.8.0

    Enrich: synchronous person and company resolution

    This release introduces Enrich, a new family of synchronous endpoints that resolve people and companies from partial information, and rebuilds the dashboard around it.

    ✨ Features

    • The Enrich Profile endpoint, POST /v2/enrich/persons, resolves a Social profile from any combination of email, first name, last name, and company domain or name; a single field is enough, and the result comes back inline (no webhook needed)
    • The Enrich Company endpoint, POST /v2/enrich/companies, resolves a company from its domain and returns the result inline
    • A new Enrich page in the dashboard lets you resolve profiles and companies interactively, and is now the default landing page after login and onboarding
    • Enrich results now include a preview of the alternative candidates evaluated during matching
    • Workspaces on trial can use the new Enrich endpoints without a paid plan

    🚀 Improvements

    • Both Enrich endpoints cost 2 credits per resolved result, and not-found lookups are free
    • Enrichment now leads the navigation, docs, and pricing pages, and replaces the legacy dashboard home
    • Enrich results render as profile cards with inline reveal, clearer error states, and validation feedback before you submit
    • The memberId field in person responses is now always a string, like publicId and linkedinUrl (previously nullable)
    • The Enrich playground opens with all input fields available and accepts URL-style domains

    🐛 Bug Fixes

    • Your credit balance now updates immediately after you run an enrichment from the dashboard
  49. v2.7.0

    Sturdier webhooks, accurate industries and stale cache billing

    Webhook polling and events get sturdier, industries follow the official Social industry taxonomy, cached records older than 90 days are now billed as a fresh lookup, and the Activities endpoints are discontinued on July 1, 2026.

    ✨ Features

    • Live endpoints now show an inline SLA disclaimer about their best-effort service level

    🚀 Improvements

    • Industry names now follow the official Social industry taxonomy
    • Polling an errored async job now returns the stored result payload instead of an empty response
    • The original request payload is now propagated to the final webhook events

    🐛 Bug Fixes

    • employeeCountRange now shows 10,000+ instead of 1 when the range has no upper bound
    • Schemeless and company Social URLs are now normalized and accepted in the person search filter
    • Upstream timeouts now return 503 instead of 408 to match the documented error contract
    • The platform status widget no longer fails to load because of a cross-origin error

    💥 Breaking Changes

    • Cached records older than 90 days are now charged as a fresh lookup instead of being served for free

    ⚠️ Deprecated

    • The Activities endpoints (person and company posts, comments, and reactions) are discontinued on July 1, 2026
    • Contact Email Finder (/v2/contact/email) was unavailable during the July 1 transition and is now available again for opted-in workspaces
  50. v2.6.0

    Refreshed interface, email preferences, and public docs

    A redesigned dashboard, a page to choose the emails you receive, and documentation you can read without logging in.

    ✨ Features

    • A new email preferences center lets you choose which lifecycle and notification emails you receive, with one-click unsubscribe
    • The API documentation is now available without logging in at docs.reversecontact.com
    • Locations across the app now show country flag emojis for faster scanning

    🚀 Improvements

    • The dashboard gets page transitions, animated KPI counters, smoother tab and filter motion, and fluid typography
    • Modals, toasts, cards, and colors now follow a single design system, and the payment success page is refreshed
    • Trial and workspace banners have a cleaner layout and styling
    • The JsonViewer adds search with data-path tracking and stays responsive on large payloads
  51. v2.5.0

    Usage insights, transparent pricing, and a multi-conversation Agent

    See exactly what each call costs, break down usage by channel, and keep multiple Agent conversations going at once.

    ✨ Features

    • The Agent now keeps multiple persistent conversations, each with a title generated automatically from your first message
    • You can filter usage and credit consumption by channel (API, Agent, dashboard), with dedicated KPI cards
    • The dashboard now shows the credit price of each endpoint, so you can see what a call costs before you run it
    • You can update your billing details from the billing overview

    🚀 Improvements

    • The quotas response now reports each key's RPM and daily limits
    • New accounts now start a trial automatically, get a default API key and see clearer pay-as-you-go copy, and onboarding no longer has a survey step

    🐛 Bug Fixes

    • The remaining-credits cards are easier to read
    • The daily rate limit now displays correctly when the gateway returns a dailyLimit object instead of a number
  52. v2.4.0

    Guided lookups, richer profiles, and a docs home

    Find the right guide faster, trigger live lookups when a profile is missing, and get fuller current-job details on every person result.

    ✨ Features

    • A documentation home at /docs helps you find the right guide and endpoint faster
    • When a profile is not in the database, you can now trigger a live lookup directly from the result
    • Lookup input fields now show format examples so you know what to paste

    🚀 Improvements

    • Person results now include full current-job details from currentPosition, including title, company, and dates
    • Usage and log dates now group by calendar day instead of a rolling 24-hour window

    🐛 Bug Fixes

    • Tall payloads in the JsonViewer now scroll inside the viewer instead of being clipped, and no longer scroll the page behind them
    • Calls to /v2/usage and other zero-cost endpoints no longer run a credit check
    • Enterprise pay-as-you-go top-ups and auto-top-up no longer apply a legacy x2 credit multiplier
  53. v2.3.0

    Agent, webhook polling & revamped API documentation

    Chat with the new Agent to run lookups in natural language, retrieve async job results from a new polling endpoint, and read the redesigned API documentation. Some of the features below are behind feature flags and roll out progressively, so your account may not show them yet: contact support if you want early access.

    ✨ Features

    • The Agent, a new AI chat interface, runs person and company lookups in natural language (private beta, feature-flagged)
    • A new webhook polling endpoint returns async job status and results without waiting for a callback (private beta, feature-flagged)
    • On the new Job Lookup page under Logs, paste a webhookId to watch an async job progress in real time, with direct links from Webhook Events (private beta, feature-flagged)
    • The redesigned support bubble has Support and Tips tabs, and shows proactive tip teasers on V2 lookup pages
    • The remaining credits cards now have a Buy credits button for faster top-ups
    • llms.txt and llms-full.txt are now published so AI agents can index the documentation

    🚀 Improvements

    • API endpoint pages are redesigned with a field list, a sticky examples panel and an embedded API playground modal
    • The docs playground is rebuilt with a shared body form, a command palette endpoint picker and more search filters
    • We rewrote the guides on authentication, the first API call, which endpoint to use, API key management, rate limits and credits, errors and retries, webhooks, data freshness and the production checklist
    • Person activities are split into three dedicated endpoint pages, and the company endpoint pages are rewritten to follow the same pattern
    • The MCP setup guide shows how to connect AI agents to the Reverse Contact API
    • The Post Reactions endpoint now handles live reactions, with clearer error messages
    • The search input placeholder is now in English across the app

    🐛 Bug Fixes

    • Bug fix bullets in the changelog are now visible in dark mode
    • Requests to custom webhook URLs are now forwarded correctly and their payload is stored
    • The async webhook docs now show the correct error callback payload structure
    • The article scroll position now resets correctly when you navigate the docs
    • The overage cost now shows the effective per-credit value for legacy accounts
    • The sync profile endpoint docs now show the correct response structure
  54. v2.2.0

    API key controls, MCP for AI agents & documentation upgrades

    Fine-tune each API key with custom rate limits, permissions, and expiry dates, and connect AI agents to your account via MCP.

    ✨ Features

    • You can configure rate limits (RPM, daily), endpoint permissions, and expiry dates for each API key from the dashboard
    • You can connect AI agents like Claude to the Reverse Contact API through MCP, with 12 intent-based tools (beta)
    • The onboarding survey can now be skipped
    • The plan summary now shows your RPM limits, with an upgrade hint when applicable

    🚀 Improvements

    • A new API key management guide covers key configuration
    • The API documentation now shows which endpoints require a paid plan
    • The API docs now mark the dailyLimit parameter as Enterprise-only
    • Upgrade prompts now link directly to the relevant documentation
    • The notification badge is easier to see in the header

    🐛 Bug Fixes

    • Fixed timezone and display issues with API key expiration dates
    • The RPM input no longer allows values above the workspace limit
    • Source labels now display correctly for activities and webhook results
    • The onboarding survey modal now scrolls on mobile
    • Trial users opening the billing portal now see a clear message instead of a 400 error
  55. v2.1.0

    Free trial, Google sign-in & new pricing tiers

    New users get a free trial, Google One Tap speeds up sign-in, and pay-as-you-go gets two new pricing tiers.

    ✨ Features

    • New users can try the platform on a free trial, with no payment required upfront
    • Google One Tap is available on the login and signup pages for faster account creation
    • Pay-as-you-go adds two pricing tiers, $100 for 5,000 credits and $250 for 15,000 credits
    • The usage page now focuses on credits, with a dynamic burn rate visualization
    • You can switch between the current month and the past month on the usage page

    🚀 Improvements

    • The trial banner is redesigned with clearer messaging and direct links to the credits tab
    • Onboarding now highlights the pay-as-you-go plan

    🐛 Bug Fixes

    • File download links no longer expire when you click enrichment results
    • The docs drawer now opens correctly from search pages
    • Person search no longer offers an unsupported email filter that could cause errors
    • Unauthenticated users can now access the public docs
    • Trial users can no longer configure auto top-up
  56. v2.0.8

    More API keys and onboarding improvements

    Create more API keys, pick your theme during onboarding, and get more accurate burn rate projections on the dashboard.

    ✨ Features

    • You can now create up to 10 API keys per workspace (previously 5)
    • A light/dark mode toggle is available during onboarding

    🚀 Improvements

    • The dashboard now uses a 7-day rolling average for more accurate burn rate projections
    • The docs sidebar now lists guides above endpoints
    • The rate limits documentation is clearer and more accurate
    • Webhook deliveries now include a webhookId field and use a structured data payload

    🐛 Bug Fixes

    • The pricing slider and premium text are now readable in light mode
    • Fixed a display issue with the loading indicator
  57. v2.0.7

    Credit monitoring and live status

    This release adds real-time credit monitoring, low-balance alerts and a live platform status indicator.

    ✨ Features

    • A new dashboard section shows your remaining credits, burn rate and requests per minute
    • You are notified automatically when your credits reach 10%, 5%, and 0%
    • The sidebar shows a live platform status indicator

    🚀 Improvements

    • The API documentation now describes the 422 WEBHOOK_URL_REQUIRED error on all async endpoints
    • Async enrichment endpoints return a 422 error when no webhook URL is configured

    🐛 Bug Fixes

    • Text on the pay-as-you-go pricing card is now readable in light mode
    • Invited users are now correctly redirected to the invitation page instead of onboarding
  58. v2.0.6

    Subscription management and reliability improvements

    You can now cancel your subscription from the billing portal, and credits are refunded automatically when an enrichment fails with a data error.

    ✨ Features

    • You can cancel your subscription from the billing portal, either immediately or at the end of the period
    • You can view the raw JSON response for any search result from the dashboard

    🚀 Improvements

    • Social URL validation now accepts names with diacritics and locale-specific subdomains
    • Credits are refunded automatically when an enrichment fails with a data error
    • The credit indicator now shows a "for 100 results" subtitle

    🐛 Bug Fixes

    • The default theme is now dark instead of following the system preference
    • Enrichment results show a smoother transition when no data is found
    • On iPad Pro, the Terms of Service accept button is now reachable with zoom enabled
    • Your color mode preference is now restored correctly on page load
    • Text, borders and chips have better contrast in light mode across the dashboard
    • Slideovers now use the correct background color in light mode
    • Auth pages no longer show white text on a white background in light mode
    • Fixed an issue that caused high backend CPU usage under sustained load
  59. v2.0.5

    Onboarding branding and dark mode fixes

    Workspace onboarding now detects your domain automatically, plus readability fixes across the dashboard.

    ✨ Features

    • Onboarding now detects your workspace domain and branding automatically

    🚀 Improvements

    • Error breakdown counts now display with thousand separators
    • API error codes no longer include internal prefixes

    🐛 Bug Fixes

    • Chart labels are now visible in dark mode
    • Enrichment result pages no longer show an inaccurate webhook delivery preview
  60. v2.0.4

    Smarter notifications and changelog improvements

    Credit alerts no longer repeat, and the changelog now highlights what you haven't read yet.

    🚀 Improvements

    • Credit alerts are now sent once per threshold instead of repeating
    • New changelog entries are highlighted so you can see what you haven't read yet

    🐛 Bug Fixes

    • Public documentation pages no longer redirect to the login screen
  61. v2.0.3

    Dashboard polish and error visibility

    Better error visibility on endpoints, dark mode readability, and small UX improvements across the dashboard.

    ✨ Features

    • Endpoint statistics now show separate 4xx and 5xx error columns for clearer debugging

    🚀 Improvements

    • Buttons show smoother loading indicators across the app
    • Chart axes are now easier to read in dark mode
    • User inputs are now trimmed automatically to prevent whitespace-related errors

    🐛 Bug Fixes

    • API key revocation and role change emails no longer have missing fields
    • The Terms of Service scroll prompt is now clearly visible before you accept
  62. v2.0.2

    UI polish and profile slug handling

    Small quality-of-life fixes across the dashboard and improved Social URL handling.

    🚀 Improvements

    • Enrichment now accepts Social profile slugs that contain Unicode characters or are percent-encoded
    • Documentation and enrichment result pages have improved responsive layouts

    🐛 Bug Fixes

    • The billing page no longer shows a daily limit for unlimited plans
    • All billing managers can now see the current plan and payment method
    • The fetch profile and company pages have clearer action labels
    • The API documentation now shows the correct webhook response format
  63. v2.0.1

    Integrations, smarter notifications, and API improvements

    New integrations with Clay and Zapier, credit usage alerts, and API reliability improvements.

    ✨ Features

    • You now receive an alert when your credit balance reaches 10%, 5%, and 0% remaining
    • A new retrocompatible endpoint supports Clay workflows
    • A new authentication endpoint lets you run Zapier automations
    • Enterprise workspaces are set up through a dedicated provisioning system

    🚀 Improvements

    • The API now accepts Social URLs with or without the https:// prefix, and both the linkedInUrl and linkedinUrl field names
    • The API documentation splits post interactions into separate Comments and Reactions pages
    • Non-billing roles now see an "access restricted" message on billing tabs instead of a broken UI

    🐛 Bug Fixes

    • The trial status no longer stays stale after you complete a Stripe checkout
    • Auto top-up no longer retries excessively when a card payment is declined
    • An unlimited rate limit now shows as no rate limit in API responses instead of 0
  64. v2.0.0

    Reverse Contact V2 platform launch

    The V2 platform is rebuilt with a new API, real-time enrichment, webhooks, billing and a redesigned dashboard.

    ✨ Features

    • API V2 ships 7 fetch endpoints (person, company, posts, comments, reactions), resolve endpoints (company from domain, person by email or names), contact email verification, check endpoints, and search with 100 results per page by default
    • Webhooks deliver results in real time instead of async polling, with an events page that offers filtering, stats, pagination and a live tail mode
    • Live enrichment shows real-time tracking, a stop button and toast notifications, and its results persist across navigation
    • Billing adds Stripe checkout, multi-account support, auto top-up, pay-as-you-go rate limits, subscription management, enterprise credit selection and a history of credit purchases
    • The dashboard adds usage analytics, a tabbed layout, stacked bar charts, an error breakdown, KPI trend comparisons and resizable table columns
    • The API documentation is rewritten with an endpoint sidebar, a split-panel layout, full-text search, public access without login, and code examples with a copy button
    • Support chat offers real-time messaging, file upload and conversation management
    • Notifications come with user preferences and email templates
    • A feature request board lets you request and vote on features, with infinite scroll
    • You can sign in with Google or GitHub through OAuth, accept the Terms of Service and change your password
    • A GDPR/CCPA data subject blocklist is enforced with HTTP 451
    • Spotlight is redesigned with global shortcuts, direct fetch from a Social URL and docs search
    • A new page detects and manages your integrations
    • You can set log retention policies at the workspace level
    • The API Playground tests endpoints live, with a credits display, line numbers and a curl preview

    💥 Breaking Changes

    • The Authorization header must use the Bearer prefix on all V2 endpoints
    • A call delivered by webhook costs 2 credits, and a synchronous call costs 1
    • Search now returns 100 results per page by default (previously 25)

    🚀 Improvements

    • Credit pricing tiers are updated
    • The brand gets new logos and a new favicon
    • Usage stats are finer-grained, at 15-minute intervals
    • You can preview and copy API keys

    ⚠️ Deprecated

    • The Watcher feature is removed