---
title: "Create Warm Intros signal"
method: POST
path: "/api/v1/signals/warm-intros"
tags: ["Signals"]
---

# Create Warm Intros signal

`POST /api/v1/signals/warm-intros`

## Request

- Base URL: `https://api.rodz.io` (the document declares no server; this is the origin it was published from)
- URL: `https://api.rodz.io/api/v1/signals/warm-intros`
- Auth: HTTP bearer

Get your team's warm intros: the people of your ICP who already worked alongside one of your people - an employee, a founder, a happy client - at the same company and at the same time, so the first message is a referral instead of a cold one. Name the people who can introduce you, or simply give your company page: we read where each of them worked and when, and deliver the former colleagues who now hold the roles you sell to. One lead per person, however many of your people know them, with every introduction path (who, where, how long together) and a strength score to start with the warmest.

**Input:** `sourceProfiles` (LinkedIn profile URLs of the people who can introduce you, up to 500; `linkedinProfileUrls`, the name the other people-based signals use, is accepted as the same field) and/or `sourceCompany` (a LinkedIn company page, usually yours: its employees become introducers; partners, agencies, freelancers, consultants, ambassadors, affiliates, advisors and board members shown on the page are left out) - at least one of the two is required; `minOverlapMonths` (minimum months worked at the same company at the same time, 3-120, defaults to 6); `includeCurrentCompanies` (also use the companies your introducers work at today; defaults to false, since their current colleagues are usually your own team or your client); `oneTimeExtractCount` (how many people to deliver once, counted in leads delivered after your filters; 0-25,000, defaults to 500; not metered by `dailyLeadLimit`); `dailyNewIntros` (true to keep delivering newly found introductions afterwards, up to `dailyLeadLimit` a day; defaults to false, which stops the signal once the one-time extract is complete); optional persona targeting (`targetPersonas`, `customPersonaTitles`); optional ICP filters (`companySizeFilters`, `locations`, `industryFilters`, `excludedLocations`, `excludedCompanySizeFilters`, `excludedIndustryFilters`, `excludedTitles`, `excludedCompanyUrls`); optional `restrictToCompanies` / `restrictToProfiles` to deliver only people at your target accounts or on your list; optional enrichment toggles (`getEmails`, `getPhones`, `getSireneData`, `detectTechnologies`, `getHomepageContent`); `webhookUrl`, `dailyLeadLimit`, `duplicateHandling`, `status`. At a company over 1,000 people a person counts only if they worked in the same function as your introducer there; over 10,000, also in the same place and for at least 12 months together. People currently working with one of your introducers (your own team) are never delivered. Each lead is billed as it is delivered, once whatever the number of paths: **3 credits per lead for the one-time extract, then 5 credits per lead for the daily new introductions**. Activating with `status: active` checks that your balance covers the one-time extract (`oneTimeExtractCount` x 3 credits) and answers 402 with the estimate when it does not.

**Output:** each person is pushed to your `webhookUrl` as an enriched person lead: the contact (name, email if enriched, phone if enriched, title, location, LinkedIn profile URL, matched persona), their employer today (company name, website, LinkedIn URL), and in `signal`: `intro_paths` - every way to be introduced, best first, each with `via` (the introducer: name, LinkedIn URL, title), `entity` (the company you share: name, LinkedIn URL, employee count), `overlap` (`start` and `end` as YYYY-MM, `months`), `same_function` and `strength` (0-100: more months together, a smaller company and the same function score higher); `intro_strength` and `intro_tier` (`strong` 70+, `likely` 50-69, `possible`) for the best path; `intro_reason`, that path in one sentence ready for a message or a CRM note; `tracked_contact` (the best introducer's LinkedIn profile URL) and `tracked_company` (the shared company's page); and the best path again as flat fields, under the names the alumni signals use, for merge tags, sheet columns and CRM fields: `introducer_name`, `introducer_title`, `affiliation_name` (the shared company), `affiliation_started` and `affiliation_ended` (YYYY-MM), `overlap_months`, `same_function`. Text that is unknown is ``, never null. The same intro fields reach Google Sheets as columns and outbound tools as merge tags (`introReason`, `introStrength`, `introTier`, `introducerName`, `introducerTitle`, `affiliationName`, `affiliationStarted`, `affiliationEnded`, `overlapMonths`, `sameFunction`). The single-signal responses report progress in `applied.oneTimeExtract` (`target`, `delivered`, `status`).

**Use cases:**
- Your team's warm intros - give your company page and get the people of your ICP who worked alongside one of your employees, with who can introduce you, where and for how long, so the first touch is a referral instead of a cold message.
- Happy-client referrals - name your best customers and reach their former colleagues who now hold the roles you sell to, then ask the client for the intro.
- Account-based prospecting - restrict to your target accounts and see, account by account, which of your people already know someone inside.
- Prioritised outreach - sort by `intro_strength` and start with the strongest paths (years together, a small company, the same function).

## Request body

- object: Create a Warm Intros signal.
  - `name` string, required: Name for this signal
  - `webhookUrl` string, nullable: Where each lead is POSTed as it is produced. Optional (since 2026-09-23): omit it when you read leads with GET /api/v1/signals/{signalId}/leads or route them to a connected CRM or sequencer with `destinations` on PATCH /api/v1/signals/{signalId}.
  - `config` WarmIntrosConfig, required: Config for warm_intros: the people your introducers worked alongside. Name the people who can introduce you — your team, your happy clients — with `sourceProfiles`, and/or give your company page with `sourceCompany` to use its employees. For each introducer we read their past companies and deliver the people matching your filters (personas, company size, location, industry, exclusions) who worked at the same company AT THE SAME TIME, for at least `minOverlapMonths`. At a company over 1,000 people the person must have worked in the same function as the introducer; over 10,000, also in the same place and for at least 12 months together. One lead per person, carrying every introduction path and a strength score; people currently working with an introducer (your own team) are never delivered. - **One-time extract** (`oneTimeExtractCount`): up to that many people, delivered as fast as they are found, not metered by `dailyLeadLimit`. 3 credits per lead, however many paths it carries. - **Daily new introductions** (`dailyNewIntros`): once the one-time part is complete, keep delivering the introductions found afterwards (a new introducer, a person who joins a searched company's alumni), under `dailyLeadLimit`. 5 credits per lead. Off, the signal stops after the one-time part.
    - `restrictToCompanies` union[]: Only deliver leads at these companies. Each entry is either a plain string — a name, a domain, or a LinkedIn company URL (e.g. `"Doctolib"`, `"payfit.com"`) — or an object `{name, domain, linkedinUrl}` (all optional, at least one required) carrying everything you know about the company; both forms can be mixed. The more identifiers you send, the more of your list is recognised. Matching is exact once punctuation and legal suffixes are ignored. Consumed on create, so it is NOT echoed back in `config`; read and edit it at `GET|POST|DELETE /api/v1/watchlists/{signalId}/companies`. **Accepted entry forms:** a company name (`"Doctolib"`); a domain or website URL (`"payfit.com"`, `"https://www.payfit.com/fr"`); its LinkedIn company page (`"https://www.linkedin.com/company/payfit"`, in any spelling: with or without `https://` and `www.`, a country subdomain, a trailing path or query; school and showcase pages too); its numeric LinkedIn company id (`"10426145"`) or a Sales Navigator company link (`"https://www.linkedin.com/sales/company/10426145"`), both read as that id; or an object `{name, domain, linkedinUrl}` combining several. Any other LinkedIn link names no company: a person's profile (`/in/`), a post or a group page is refused with 422 `INVALID_COMPANY_ENTRY`, which lists each offending entry, and nothing is saved.
      - union
        - string
        - WatchedCompanyObject: One company, with everything you know about it. The object form of a company-list entry. Send every identifier you have: each one is another way to recognise the company, and together they separate two companies that share a name.
          - `name` string, nullable: The company name as you know it, e.g. `"PayFit"`.
          - `domain` string, nullable: The company website, e.g. `"payfit.com"`. A full URL is accepted and reduced to its domain.
          - `linkedinUrl` string, nullable: The company LinkedIn page, e.g. `"https://www.linkedin.com/company/payfit"`. The most precise identifier of the three. A Sales Navigator company link (`/sales/company/<id>`) or the numeric LinkedIn company id is accepted too, and so are school and showcase pages; any other LinkedIn link (a person's profile, a post) is refused with 422 `INVALID_COMPANY_ENTRY`.
    - `restrictToProfiles` string[]: Only deliver leads for these people — LinkedIn profile URLs. Consumed on create like `restrictToCompanies`; read and edit it at `GET|POST|DELETE /api/v1/watchlists/{signalId}/profiles`.
    - `companySizeFilters` string[]: Company size filter. Values: "Self-employed", "1-10", "11-50", "51-200", "201-500", "501-1000", "1001-5000", "5001-10000", "10001+". A company qualifies when its declared size band OR its employee count falls in a selected band; a count within 10% of a band edge counts for both neighbouring bands, and a company declaring 501+ employees is never placed in a smaller band by a low count. The delivered `employee_range` / `employee_count` are the pair the lead was judged on.
    - `targetPersonas` string[]: Filter by job-role persona (decision-maker targeting). Accepted values are the persona display names listed here; see GET /v1/personas for the full taxonomy with department, seniority and EN/FR job titles. Titles are matched by meaning, not as text: `targetPersonas` and `customPersonaTitles` form ONE list and a person qualifies by matching ANY entry. Each entry is read as a function plus a seniority level, in any language ("Directeur Commercial" satisfies "Sales Director", "Président" or "Directeur général" a CEO target), and a role at that level or above qualifies: a Director or Head target refuses a plain Manager; a founder running the company counts as CEO. A custom title carries the level its own words state; one with no level word accepts every level. Valid values: "Account Executive", "Account Manager", "Accountant", "Agency Director", "Business Analyst", "Business Developer", "Business Unit Director", "Buyer", "CEO", "CIO", "CSR Director", "CSR Manager", "CTO", "Communication Director", "Community Manager", "Content Manager", "Copywriter", "Creative Director", "Customer Success Director", "Customer Success Manager", "Customer Support Agent", "Cybersecurity Director", "Data Analyst", "Data Director", "Designer", "DevOps Engineer", "Digital Director", "Digital Project Manager", "Ecommerce Director", "Ecommerce Manager", "Engineering Manager", "Engineering Office Director", "Ergonomist", "Executive Assistant", "Finance Director", "Finance Manager", "GTM Director", "GTM Manager", "Growth Director", "Growth Manager", "HR Assistant", "HR Development", "HR Director", "Happiness Manager", "IT Project Manager", "Industrial Director", "Innovation Director", "Lab Manager", "Legal Counsel", "Legal Director", "Legal Manager", "Logistics Coordinator", "Logistics Director", "M&A Director", "Marketing Director", "Marketing Manager", "Office Manager", "Omnichannel Director", "Operations Director", "Operations Manager", "Payroll Manager", "Payroll Officer", "Procurement Specialist", "Product Director", "Product Manager", "Project Manager", "Purchasing Director", "Purchasing Manager", "Quality Health and Safety", "Quality Manager", "Quality Technician", "R&D Director", "R&D Engineer", "R&D Manager", "Recruitment Manager", "Research Director", "Research Scientist", "SDR / BDR", "Sales Director", "Sales Manager", "Social Media Director", "Social Media Manager", "Software Engineer", "Supply Chain Director", "Support Director", "Support Manager", "Talent Acquisition", "Technical Director", "Training Director", "Training Officer", "Transformation Director", "UX Designer", "Warehouse Director", "Web Developer"
    - `locationFilter` string: Filter by location (e.g. 'France', 'Paris')
    - `locations` string[]: Target locations (e.g. ['France', 'Paris']). Applies to the PERSON's LinkedIn profile location, not the company's headquarters: a person in Paris working for a company headquartered abroad matches France. One matching entry is enough. Only fundraising and M&A signals filter on headquarters (`hqLocations`).
    - `excludedLocations` string[]: Locations to exclude from results
    - `industryFilters` string[]: Only return companies in these LinkedIn industries. Two kinds of value work: a single industry (e.g. 'Software Development') or a broad category that covers many (e.g. 'Technology and Media'). Matching is exact, so a near-miss like 'Software development' matches nothing and the signal quietly produces no leads. Empty = all industries. Call GET /api/v1/signals/industries for the valid list.
    - `companyTypeFilters` string[]: Only return companies of these legal/ownership types. Valid values: 'Privately Held', 'Public Company', 'Nonprofit', 'Educational Institution', 'Partnership', 'Sole Proprietorship', 'Government Agency', 'Self-Employed'. Empty = all types. A company whose type is unknown is KEPT, not dropped, so this narrows results only where the type is known.
    - `excludedIndustryFilters` string[]: LinkedIn industries to exclude. Same vocabulary as `industryFilters` (single industry or category name); an unrecognised value excludes nothing. Call GET /api/v1/signals/industries for the valid list.
    - `excludedCompanyUrls` string[]: LinkedIn company URLs to exclude from results
    - `excludedCompanySizeFilters` string[]: Company sizes to exclude from results
    - `customPersonaTitles` string[]: Custom free-text job titles to match in addition to predefined personas. Titles are matched by meaning, not as text: `targetPersonas` and `customPersonaTitles` form ONE list and a person qualifies by matching ANY entry. Each entry is read as a function plus a seniority level, in any language ("Directeur Commercial" satisfies "Sales Director", "Président" or "Directeur général" a CEO target), and a role at that level or above qualifies: a Director or Head target refuses a plain Manager; a founder running the company counts as CEO. A custom title carries the level its own words state; one with no level word accepts every level.
    - `excludedTitles` string[]: Job titles to exclude. Matched as whole words, case- and accent-insensitive: "intern" excludes "Intern" and "Interns", never "International"; a phrase matches the same words in the same order; end a word with * to match a prefix ("infirmi*" excludes "Infirmière").
    - `getEmails` boolean: Enable email enrichment (0.5 credit/lead, charged only on success)
    - `getPhones` boolean: Enable phone enrichment (5 credits/lead, charged only on success)
    - `getSireneData` boolean: Enrich with French SIRENE registry data (1 credit/lead, charged only on success)
    - `detectTechnologies` boolean: Detect technologies used on the company's public website (1 credit/lead, charged only on success)
    - `getHomepageContent` boolean: Fetch the company's homepage content (1 credit/lead, charged only on success)
    - `sourceProfiles` string[]: LinkedIn profile URLs of the people who can introduce you (your team, your happy clients), up to 500. `linkedinProfileUrls` is accepted as the same field.
    - `sourceCompany` string, nullable: Optional. A LinkedIn company page (usually yours) whose employees become introducers. Only real employees are kept: partners, agencies, freelancers, consultants, ambassadors, affiliates, advisors and board members shown on the page are left out. Together with `sourceProfiles`, at most 500 introducers.
    - `minOverlapMonths` integer: Minimum number of months a person must have worked at the same company at the same time as an introducer.
    - `includeCurrentCompanies` boolean: Also use the companies your introducers work at TODAY. Off by default: their current colleagues are usually your own team or your client.
    - `oneTimeExtractCount` integer: How many people to deliver, once, counted in leads delivered after your filters. Not metered by `dailyLeadLimit`; billed 3 credits per lead. `0` skips this part.
    - `dailyNewIntros` boolean: After the one-time extract, keep delivering newly found introductions, up to `dailyLeadLimit` a day, billed 5 credits per lead. `false` stops the signal once the one-time extract is complete.
  - `dailyLeadLimit` integer: Max leads per day for the daily part (`dailyNewIntros`). The one-time extract is bounded by `oneTimeExtractCount` instead.
  - `duplicateHandling` string, nullable: Duplicate lead handling: 'refuse' (never accept), 'allow' (always accept), or 'Xd' (accept a duplicate after X days, e.g. '7d'). Optional. Whatever the setting, one person found through several introducers is ONE lead.
  - `status` string, nullable: Initial status (default: draft). Accepts draft, active, or paused. If active, the config is validated and the balance must cover the one-time extract (`oneTimeExtractCount` x the price per lead), otherwise 402.

## Response `201`

Signal created

- object: A single signal: create, GET /signals/{id} and PATCH. It carries two fields the list does not. A customer reads their stored `config` back and assumes it is the one in force; on `job_changes` three fields were not (`maxJobAgeDays` 90 stored / 180 applied, `publishedDate` stored and never read, `allowUnknownCompanySize` true and applied false), with nothing anywhere saying so.
  - `id` string, required
  - `name` string, required
  - `status` string, required
  - `signalTypeId` string, required
  - `webhookUrl` string, nullable
  - `config` object, nullable: Signal-specific configuration
  - `dailyLeadLimit` integer, required
  - `leadsToday` integer, required
  - `leadsTotal` integer, required
  - `duplicateHandling` string, required: Duplicate lead handling: 'refuse' (never accept), 'allow' (always accept), or 'Xd' (accept after X days)
  - `createdAt` string, required
  - `updatedAt` string, required
  - `lastLeadAt` string, nullable
  - `applied` object: Settings this signal runs with where they differ from `config`, keyed like `config`. `{}` when every stored setting is applied as stored. A `null` value means the stored field is not used by this signal. Every entry is explained by a warning naming the same field.
  - `warnings` SignalConfigWarning[]: What you should know about this configuration: values applied differently from what is stored, and filter values that are probably not what you meant. Computed from the stored configuration, so they appear on GET as well as on create and update - a signal created before these fields existed reports them too. Never an error: the signal is created or updated regardless.
    - `code` string, required: Stable identifier. `VALUE_BELOW_MINIMUM` (a value under the documented minimum is applied as the minimum), `VALUE_ADJUSTED` (a value the search cannot use as written is applied as the nearest one it can, e.g. a `publishedDate` of 14d searches 15 days), `FIELD_NOT_USED` (a stored field this signal ignores), `UNKNOWN_SIZE_NOT_ACCEPTED`, `CATEGORY_WITH_OWN_INDUSTRIES` (a category listed together with industries it already covers), `UNKNOWN_INDUSTRY` (a value matching no published industry), `UNKNOWN_FIELD` (a config key this signal does not have), `UNKNOWN_TOOL` / `UNKNOWN_CATEGORY` (a tracked tool or category name that is not in the catalog of GET /api/v1/signals/technographic-tools; on a GMB Discovery signal, a `category` that is not a Google Maps category, so the signal matches no business), `UNKNOWN_CITY` (GMB Discovery: cities in `cities` that are not in the signal's `country` — `values` lists them; when none is, the signal matches no business until `cities` or `country` is changed; never an error, nothing stored is changed), `NO_FOLLOWER_SOURCE` (a Company Followers signal with no follower extraction set up yet, so it produces nothing), `DAILY_LIMIT_PENDING` (a job signal's `dailyLeadLimit` changed less than a day after its search was last rebuilt: the search keeps its previous volume until the time in `values`, while the leads delivered per day already follow the new limit), `WATCHED_LIST_EMPTY` (a watchlist reading its list from your CRM holds no company or person — the CRM is not connected, the list could not be read, or it is empty — so it produces nothing until the list fills; cleared on its own once it does), `QUOTE_REQUIRED` (an `active` Company Followers signal in `icp` mode whose follower extraction has not been priced and accepted, so it produces nothing yet: `values[0]` says why — `not_requested`, `awaiting_acceptance` (then `values[1]` is the quoted price in credits), `expired`, `declined`, `failed`, or, after an automatic quote, `auto_quote_refused` / `auto_accept_refused` (then `values[1]` is the error code); the message names the call to make — see POST /api/v1/signals/{signal_id}/extraction/quote; cleared once a quote is accepted), `WATCHED_LIST_IMPORTING` (a company watchlist whose HubSpot list is still being read: companies are added as they arrive; `values[0]` is the number of members read so far, `values[1]` the list id; cleared once the whole list is read), `MATCHING_CONFIRMATION_REQUIRED` (a New Decision Maker or Account watchlist whose companies must be matched to their LinkedIn company page before they can be watched, at a cost above the amount we match without asking: nothing is matched or charged until you confirm with POST /api/v1/watchlists/{signalId}/companies/confirm-matching; `values[0]` is the maximum cost, `values[1]` the number of companies to match; cleared once confirmed), `DESTINATION_UNAVAILABLE` (the outreach campaign this signal sends leads to cannot receive them — not found, archived, ended or without any step — so leads are not reaching it: `values[0]` is `campaign_not_found`, `campaign_archived`, `campaign_ended` or `campaign_no_steps`; cleared once a lead reaches the campaign or the signal's destination is changed), `DESTINATION_NEEDS_CONTACT` (the signal delivers companies with no contact and a destination in `destinations` only accepts people, so none of its leads reach that destination), `DESTINATION_SKIPS_COMPANY_ONLY` (some leads are companies with no contact, e.g. no one matching your personas was found, and a destination only accepts people), `DESTINATION_NEEDS_EMAIL` (a destination only accepts leads with an email and `getEmails` is off) — for these three, `values[0]` is the destination key in `destinations`, the next values are the fixes: `add_persona` (turn on `enablePersonaTargeting` with `targetPersonas`), `require_contact` (turn on `requireContact`), `add_company_destination` (add a destination that accepts companies: `hubspot`, `pipedrive`, `googleSheets` or `customWebhook`), `enable_email_finding` (turn on `getEmails`); severity `warning` when those leads reach no destination at all, `info` when another destination still receives them; never blocks the save; `COMPANY_PAGE_UNAVAILABLE` (a Company Followers signal whose extraction stopped because a page in `linkedinCompanyUrls` does not exist on LinkedIn: nothing was extracted or charged; `values` lists the pages; replace them with PATCH /api/v1/signals/{signal_id}, then start the extraction again; cleared once the pages are replaced or a new extraction runs), `COMPANY_PAGE_UNVERIFIED` (create / PATCH responses only: a page in `linkedinCompanyUrls` could not be verified at that moment and was accepted; `values` lists them — if one does not exist, its extraction stops without using credits and the signal carries `COMPANY_PAGE_UNAVAILABLE`).
    - `severity` string, required: `info`: nothing to do. `warning`: probably not what you meant.
    - `field` string, required: The `config` key concerned, camelCase (or `dailyLeadLimit`).
    - `message` string, required
    - `values` string[]: The offending values, when there are some.

## Other responses

- `400`: Invalid request: missing required config (Introducers (sourceProfiles or sourceCompany)), or an unrecognized location
- `401`: Missing or invalid API key
- `402`: Account has no credits
- `422`: Validation Error
- `429`: Rate limit exceeded. Retry after the number of seconds in the `Retry-After` response header.

---

[API](https://skmtc.dev/rodz/apis/rodz-api.md) · [All operations](https://skmtc.dev/rodz/apis/rodz-api/llms.txt) · [OpenAPI document](https://skmtc.dev/rodz/apis/rodz-api/revisions/3003505ff896?raw)