Documentation
fyndr connector for Claude
The fyndr connector is a remote MCP server that lets you work with your fyndr recruiting data from Claude — search jobs, review candidates, and draft recruiting content in a conversation. It is read-only: the connector can never create, edit, or delete anything in fyndr.
Prerequisites
- A fyndr account. Sign up at talent.thefyndr.com.
- Job search (
search_jobs,get_job) works with any account. - Candidate and AI tools require a company account— your email must be on a company’s allowed list, which a fyndr admin configures.
- A Claude plan that supports custom connectors.
Connecting in Claude
- In Claude, open Settings → Connectors and choose Add custom connector.
- Enter the server URL:
https://talent.thefyndr.com/api/mcp - Claude registers itself automatically and opens the fyndr sign-in page. Sign in with Google or your fyndr credentials.
- Review the consent screen — it lists exactly what the connector can access — and click Approve. You are returned to Claude with the connector active.
- The connector’s tools are now available in your conversations.
Authentication
The connector uses OAuth 2.0 — authorization code flow with PKCE (S256) and dynamic client registration (RFC 7591). Discovery metadata is published at /.well-known/oauth-authorization-server and /.well-known/oauth-protected-resource. Access tokens are JWTs that expire after one hour; there are no refresh tokens. Authorization codes are single-use and expire after five minutes. To revoke access, remove the connector in Claude’s connector settings.
Tool reference
search_jobsread-onlyAny connected accountSearch and list job postings. Returns title, location, type, salary range, and deadline.
| query | optional — match against the job title |
| type | optional — full-time | part-time | contract | remote |
| location | optional — partial match on location |
| status | optional — open (default) | draft | closed |
| limit | optional — 1–50, default 10 |
get_jobread-onlyAny connected accountGet the full details of one job posting, including description and requirements.
| job_id | required — the job's ID, as returned by search_jobs |
list_candidatesread-onlyCompany & admin accounts onlyList job-seeker profiles that have completed onboarding. Returns name, location, and top skills. Contact details are not exposed through the connector.
| skills_filter | optional — comma-separated skills, e.g. "React,Node.js" |
| limit | optional — 1–50, default 20 |
match_candidates_to_jobread-onlyCompany & admin accounts onlyRank job-seeker profiles against a specific job by skill overlap. Returns the top matches with a score, their skills, and their most recent role.
| job_id | required — the job to match against |
| top_n | optional — 1–20, default 5 |
generate_job_descriptionread-onlyCompany & admin accounts onlyDraft a plain-text job description with AI. Returns text only — nothing is saved to fyndr.
| title | required — the job title, e.g. "Senior React Engineer" |
| context | optional — key requirements, tech stack, or notes to include |
generate_interview_questionsread-onlyCompany & admin accounts onlyDraft 12 tailored interview questions for a role (technical, behavioral, problem-solving, culture-fit, scenario). Returns text only — nothing is saved to fyndr.
| job_id | optional — pull the role from an existing job |
| job_title | optional — used if job_id is not provided |
| job_description | optional — used if job_id is not provided |
| requirements | optional — list of requirement strings |
Data handling
- The connector calls fyndr’s own first-party API and database. Job and candidate data is your organization’s data.
generate_job_descriptionandgenerate_interview_questionssend the title and context you provide to OpenAI to produce text. Nothing is written back to fyndr.- Candidate contact details (email, phone) are never returned through the connector — look them up in the fyndr dashboard.
- The connector does not read your Claude conversations, memory, or files.
- Full details are in the Privacy Policy.
Privacy policy
The full policy is at talent.thefyndr.com/privacy. Summary as it applies to this connector:
- Data collected. Your fyndr account identity (name, email) via OAuth; the tool name and parameters of each request; OAuth client registrations and short-lived authorization codes/tokens.
- Use and storage.Requests are used only to serve that call. Job and candidate data stays in fyndr’s MongoDB Atlas database (hosted in the US); the app runs on Vercel. Data is encrypted in transit (TLS) and at rest.
- Third parties.
generate_job_descriptionandgenerate_interview_questionssend the title and context you supply to OpenAI (as an API provider that does not train on the data) to produce text. Tool results are returned to your Claude client and processed by Anthropic under its terms. No other sharing. - Retention. Authorization codes are deleted at expiry (5 min) or on use; access tokens expire after 1 hour and there are no refresh tokens; unused OAuth client registrations are deleted after 90 days. Connector requests are not stored beyond standard short-term server logs.
- Not accessed. Your Claude conversation history, memory, or files. Candidate contact details (email, phone) are never returned by the connector.
- Contact. Data access, export, or deletion requests: info@thefyndr.com.
Rate limits
AI generation tools are limited to roughly 30 calls per hour per user. Read tools are not rate-limited beyond normal fair-use protection. Result sets are capped (job and candidate lists default to 10–20 items, max 50).
Support
Email info@thefyndr.com with connector issues. Include the tool name and roughly when the request failed.