Back to Home

Figlime MCP Connector

Connect Figlime to Claude, ChatGPT, Cursor, or any MCP client — and work with your jobs, screening questions, and candidate responses in chat, under your existing permissions.

What it does

Figlime is a recruiting screening platform. This connector lets an AI assistant read a workspace’s jobs, screening questions, and candidate / screening responses, and propose draftscreening questions for a human to review. All recruiting data is read from Figlime’s synced ATS mirror — it is never fetched live from your ATS through this connector.

Requirements

  • A Figlime workspace with at least one job.
  • A Figlime account that is a member of that workspace (sign-in is via Google). Connecting from the workspace settings requires an admin role.
  • An MCP-capable client (e.g. Claude, ChatGPT, Cursor, VS Code).

Connect

The server is a remote, Streamable-HTTP MCP endpoint:

https://figlime.rise2.ai/mcp

  1. In your MCP client, add a custom / remote connector and paste the URL above.
  2. You’ll be redirected to Figlime to sign in (Google) and review the request. This is a standard OAuth 2.1 flow with PKCE — the client receives a scoped access token, never your password.
  3. On the consent screen, choose which workspace(s) the connection may access, the permission scope (read-only, or read + draft questions), and a role ceiling. The connection can only ever do the intersection of what you granted and what your account is allowed to do.
  4. Approve, and the tools below become available in your client.

You can review or revoke a connection at any time from Config → Integrations → MCP in your Figlime workspace.

Tools

ToolAccessWhat it does
list_workspacesReadLists the workspaces this connection can access. Use a slug as `company` for the other tools (omit it when there's exactly one).
list_jobsReadOpen jobs (id, name, department) for a workspace.
get_candidate_countsReadExact candidate (application) count per job, including closed jobs. The cheap, correct way to answer “how many candidates per job”.
get_workspace_snapshotReadEvery open job with its screening questions, sessions, and a capped sample of candidates plus the total count per job.
get_job_questionsReadThe screening questions configured for one job.
get_job_responsesReadScreening sessions and candidates for one job, with the total candidate count.
update_job_questionsDraft writeReplace a job's screening questions as a DRAFT. Drafts are never shown to candidates until a human re-publishes them; every question is compliance-checked first.

Read tools run without per-call confirmation; update_job_questions always prompts, and only ever writes a draft a human must re-publish.

Example prompts

  • How many candidates are there for each open job? — uses get_candidate_counts
  • Summarize the screening responses for the Senior Backend Engineer role and flag the strongest two candidates. — uses get_job_responses
  • Show the screening questions for the Account Executive job, then draft a tighter set focused on quota attainment (save as a draft). — uses get_job_questions, update_job_questions

Permissions & security

  • Your roles still apply. Every call is re-checked against your live Figlime role at request time, intersected with what you granted at consent. Revoking or downgrading your access immediately disables the connection — no stale grant outlives it.
  • Writes are advisory. The connector cannot publish candidate-facing changes or make hiring decisions. update_job_questions saves a draft only.
  • Scoped data. The connector reads only recruiting data in the workspaces you authorized — never your chat history, files, or other workspaces.

Support

Questions or issues: support@rise2.ai.