MCP server

Refine Job Search

co.getrefine/refine

Checked 26 Oct 2026, 08:29 UTC · Connected to it and listed its tools

Job search management: applications, interviews, offers, recruiter email, and company openings.

Website

At a glance

Type

MCP server

Publisher

getrefine.co (verified by the MCP Registry)

Status

Working at last check

Tools

62

Token cost

About 13,467 tokens per request

Sets up in

Claude Code, Claude Desktop, Codex, Gemini CLI, Cursor, VS Code, Antigravity, ChatGPT, OpenAI API, Grok

Last checked

26 Oct 2026, 08:29 UTC

Connect it to your agent

Add to CursorAdd to VS CodeVS Code Insiders

Opens the app, which shows the server and asks you to confirm. For other apps, copy the setup below.

Claude Code

claude mcp add --transport http refine https://mcp.getrefine.co/mcp

Claude Desktop

Settings → Connectors → Add custom connector, and paste the URL.

Codex (config.toml)

[mcp_servers.refine]
url = "https://mcp.getrefine.co/mcp"

Gemini CLI

{"mcpServers":{"refine":{"httpUrl":"https://mcp.getrefine.co/mcp"}}}

Cursor

{"mcpServers":{"refine":{"url":"https://mcp.getrefine.co/mcp"}}}

VS Code

{"servers":{"refine":{"type":"http","url":"https://mcp.getrefine.co/mcp"}}}

Antigravity

agy mcp add refine https://mcp.getrefine.co/mcp

ChatGPT

Settings → Apps & Connectors → Advanced settings: turn on Developer mode, then Create a connector and paste the URL. Which plans can do this varies.

OpenAI API (Responses tools)

{"type":"mcp","server_label":"refine","server_url":"https://mcp.getrefine.co/mcp","require_approval":"always"}

Grok (xAI API tools)

{"type":"mcp","server_label":"refine","server_url":"https://mcp.getrefine.co/mcp"}

Tools

• account.agenda ~54 tokens
  Get upcoming interviews and action items
• companies.submissions.create ~183 tokens
  Submit a missing company using its public HTTPS careers URL, supplied or confirmed by the user. Refine detects supported sources and processes openings asynchronously. This adds to the shared index, not to Following. Use companies.submissions.get for progress; follow separately once company.id is av
• companies.submissions.get ~145 tokens
  Read company submission progress. queued/detecting/processing means wait; ready means search again using the company name. unsupported means the careers source is not supported; failed means processing failed. Do not claim the company was added or invent jobs while it is still working.
• following.digest.get ~91 tokens
  Read the user-wide Following email digest schedule. data is null if no schedule is configured. Company match filters are managed separately with following.update.
• following.digest.update ~172 tokens
  Set the user-wide Following email digest schedule after the user requests it. Requires Plus or Student. For a new schedule supply frequency, hour (0–23), timezone (IANA), and weekday (Sunday=0) when weekly. Existing schedules can be patched.
• following.follow ~854 tokens
  Follow an indexed company for future matching openings and company recommendations. Get company_id from Search, a ready submission, or a recommendation candidate_company.id. Specify at least one match; [{}] follows all new openings. This does not submit an unknown careers page or collect historical
• following.list ~101 tokens
  List all followed companies, their filters, state, and unread counts. Returns the full list without pagination. Following delivers future matching openings; use Search to browse jobs already in the index.
• following.openings.list ~286 tokens
  Read matching openings from followed companies. tab defaults to unread. Use the row id for marking read and nested opening.id for opening details or adding to the board. meta.unread_count/read_count cover the selected follow or all companies before view filters; meta.filtered_count covers the curren
• following.openings.mark_all_read ~165 tokens
  Mark all unread openings as read for one follow, or every followed company if follow_id is omitted. Includes every page and openings hidden by view filters. Confirm the scope with the user first. The API returns no affected count; do not invent one.
• following.openings.mark_read ~95 tokens
  Mark one Following row as read. Use the row id from following.openings.list, not nested opening.id.
• following.pause ~115 tokens
  Pause future matching openings for this follow without deleting it. This does not stop shared company indexing or pause other follows.
• following.resume ~86 tokens
  Resume a paused company follow.
• following.unfollow ~143 tokens
  Unfollow a company and remove its Following feed and recommendation source. Confirm with the user first. Board jobs remain. To pause temporarily instead, use following.pause.
• following.update ~840 tokens
  Replace all match filters for a followed company. Other follow settings are unchanged. Pass [{}] for all new openings. Pending means the new filters are being applied; old unread counts may change.
• inbox.acknowledge ~158 tokens
  Keep auto-applied inbox items as-is and finalize the automatic action. Confirm the specific items with the user first.
• inbox.apply_suggestion ~151 tokens
  Apply an inbox item's suggested action, such as adding an application to the board or recording a rejection. Confirm the specific action with the user first.
• inbox.decline Sends messages · ~250 tokens
  Queue decline replies to recruiter outreach. Depending on the user's preference this SENDS email from their address. Only call after the user explicitly approved declining these specific conversations.
• inbox.dismiss ~149 tokens
  Dismiss inbox items you are not acting on (also marks the mail read in Gmail)
• inbox.draft_reply Sends messages · ~137 tokens
  Generate an AI reply draft for an inbox item. Returns HTML for review; it does not send anything.
• inbox.get ~80 tokens
  Get one inbox item with its related job and application
• inbox.get_thread ~79 tokens
  Fetch the full email thread for an inbox item
• inbox.list ~315 tokens
  List email conversations Refine matched to your job search (applications, interviews, recruiter outreach)
• inbox.mark_read ~77 tokens
  Clear an inbox item's unread-reply flag
• inbox.refresh ~60 tokens
  Check Gmail for new mail now; returns per-account sync status
• inbox.send_reply Sends messages · ~230 tokens
  Send an email reply through the user's Gmail. This emails a real person from the user's address. Only call after the user has seen the exact message body and explicitly approved sending it.
• inbox.sync_status ~68 tokens
  Per-account Gmail synchronization status, including current progress and last sync time
• inbox.undo ~142 tokens
  Undo an auto-applied inbox action. This can delete a board record that the automatic action created, so confirm with the user first.
• inbox.unsubscribe ~182 tokens
  Unsubscribe the user from this sender's mailing list via one-click list-unsubscribe. This changes the user's real email subscriptions; confirm with them before calling. Use it for unwanted bulk senders, not for a person you can simply decline.
• integrations.connect_google ~127 tokens
  Get a link for the user to connect Gmail to Refine or approve a missing Gmail permission. Refine selects the permission needed for the connected account. The connection happens in the browser with no Refine sign-in required. Share the link only with the owner of this Refine account.
• integrations.list ~76 tokens
  List connected Google accounts and their grant status. To connect a new account, use integrations.connect_google.
• jobs.add ~211 tokens
  Add a new job to your journey
• jobs.archive ~88 tokens
  Hide a job from the active board without recording a rejection or closing the application
• jobs.delete Deletes data · ~154 tokens
  Permanently delete a job and its application from the board. Confirm with the user first; it cannot be undone. To hide a job without recording a rejection or closure, use jobs.archive instead.
• jobs.get ~70 tokens
  Get details of a specific job
• jobs.interviews.add ~267 tokens
  Log an interview for a job
• jobs.interviews.delete Deletes data · ~150 tokens
  Permanently delete an interview from a job. Confirm with the user first; it cannot be undone.
• jobs.interviews.update ~664 tokens
  Update an existing interview
• jobs.list ~102 tokens
  List jobs in your journey
• jobs.mark_applied ~70 tokens
  Mark a job as applied
• jobs.offer.accept ~149 tokens
  Accept the offer for a job. This is a major state change that closes out the search for this role; confirm with the user before calling. It does not email the employer.
• jobs.offer.add_negotiation ~340 tokens
  Log a negotiation event on an offer (counter-offer, update, etc.)
• jobs.offer.create ~656 tokens
  Record a job offer with compensation details
• jobs.offer.decline ~133 tokens
  Decline the offer for a job, a major state change; confirm with the user first. It does not email the employer.
• jobs.offer.delete Deletes data · ~142 tokens
  Permanently delete the offer recorded for a job (the job stays). Use only to fix a mistake, and confirm with the user first; it cannot be undone.
• jobs.offer.delete_negotiation Deletes data · ~148 tokens
  Remove a negotiation event from an offer
• jobs.offer.update ~654 tokens
  Update the offer recorded for a job
• jobs.refresh ~98 tokens
  Queue a job posting to be re-scraped. The returned job may remain pending; use jobs.get later to inspect the refreshed state.
• jobs.unarchive ~75 tokens
  Restore an archived job to the active board
• jobs.update ~246 tokens
  Update a job in your journey
• openings.add_to_board Sends messages · ~134 tokens
  Add an indexed opening to the board after the user chooses it. Uses opening.id, not the Following row ID. Returns the existing board job if already added. This does not apply to the employer or send email; it also marks any matching Following item as read.
• openings.get ~100 tokens
  Read an indexed opening, including its full description and qualifications. Use opening.id from Search or Following, not a follow-opening row ID.
• persona.get ~80 tokens
  Get the user's persona: inferred job-search preferences and traits with an understanding score. Use it to tailor advice and searches.
• recommendations.block ~94 tokens
  Dismiss a recommended company and block it from future recommendations
• recommendations.dismiss ~111 tokens
  Dismiss a recommended company
• recommendations.feedback ~282 tokens
  Record recommendation feedback to improve future suggestions. The historical feedback tag "watched" records that a company was followed; it does not create a follow. Use following.follow first. Dismiss/block/restore actions are separate tools.
• recommendations.get ~90 tokens
  Get one recommended company with sources and persona fit
• recommendations.list ~208 tokens
  Discover companies recommended from the companies you follow. Optionally filter by source_company_id (company.id from following.list, not a follow ID). Returns one flat list plus per-source counts. To follow a recommendation, use following.follow with candidate_company.id and your chosen filters.
• recommendations.restore ~88 tokens
  Restore a dismissed or blocked recommendation
• search.company_openings ~184 tokens
  Read all matching jobs for a company in a ready search, not just its three preview jobs. Continue with data.next_page until null. Use company.id from the search results.
• search.create ~928 tokens
  Search indexed jobs or companies. Supply input OR query, not both. Natural-language searches return a search ID and processing state: use search.get later, without rapid polling. Keep the ID to paginate or update this same search. Empty query {} is All openings, currently sorted by newest without pe
• search.get ~231 tokens
  Read a search and its results, switch jobs/companies view, or fetch results.next_page. Pending/interpreting/processing means it is still working: do not create another search or poll rapidly. If missing_company_name is set, ask for its careers URL and use companies.submissions.create. Company rows c
• search.update ~909 tokens
  Update the existing search instead of creating another. Supply input OR the complete replacement query; omitted filters are cleared. Wait until the current search is ready or failed before updating it.

Browse

• Email

Related MCP servers

• jobctl
  Job application tracker for developers - AI agents write over MCP, you review in a dashboard.
• Yardstick ATS
  Manage hiring in Yardstick's ATS: jobs, candidates, applications, interviews, and scorecards.
• GigHQ.ai
  Job-search tracker for AI agents: applications, stale follow-ups, skills gaps, H-1B lookup.
• Recruitee MCP by usefulapi
  Search Recruitee candidates, jobs, pipelines and interviews, and add notes, tags and tasks.
• Jobs by Workable
  Jobs by Workable MCP server. Access over 150K jobs & thousands of companies listed on the job board.
• Job Spotter
  Search current US and UK job listings by role and place, title matches first.
• Dreamwork
  Job search, application materials, and application tracking for AI assistants.
• GetMattered
  Auto-Apply That Ends on an Interview. Job matches, per-role resumes, apply, approve, track receipts.

Badge for your README

[![AptLynx](https://aptlynx.net/badge/co.getrefine/refine.svg)](https://aptlynx.net/servers/co.getrefine/refine)

Listing text is written by the maker; AptLynx lists tools and does not vouch for them. Find another tool · Browse by topic · For agents