# MCP for agents

The hosted Trawle MCP server gives AI assistants tools to search jobs, resolve places and roles, and run a daily job search. Coming soon.

> **Preview.** The MCP server is being built and is not live yet. Names and details below may change before launch.

The Trawle MCP server will let AI assistants such as Claude and ChatGPT search Trawle directly. It will be hosted at:

```text
https://trawle.com/mcp
```

It uses the same data and the same `JobQuery` as the [API](https://developers.trawle.com/api), and the same rules: data refreshes daily, results link to the employer's posting, and "Data by Trawle" should be shown with them.

## Tools

| Tool               | Does                                                                          | API equivalent                                  |
| ------------------ | ----------------------------------------------------------------------------- | ----------------------------------------------- |
| `search_jobs`      | Search open jobs with a `JobQuery`; supports `postedSince` for new-only runs. | [`POST /v1/jobs/search`](https://developers.trawle.com/api/search-jobs)      |
| `get_job`          | One job, open or closed, with its history.                                    | [`GET /v1/jobs/{id}`](https://developers.trawle.com/api/get-job)             |
| `check_status`     | Whether jobs are still open, for a list the user saved earlier.               | [`POST /v1/jobs/status`](https://developers.trawle.com/api/check-job-status) |
| `resolve_location` | Place names to ids for `places`.                                              | [`GET /v1/places`](https://developers.trawle.com/api/resolve-location)       |
| `find_occupation`  | Job titles to role slugs for `roles`.                                         | [`GET /v1/roles`](https://developers.trawle.com/api/find-roles)              |
| `get_company`      | A company and its hiring activity.                                            | [`GET /v1/companies/{slug}`](https://developers.trawle.com/api/get-company)  |
| `search_companies` | Company names to slugs for `companies` and `get_company`.                     | [`GET /v1/companies`](https://developers.trawle.com/api/search-companies)    |
| `list_families`    | Job families for `families`.                                                  | [`GET /v1/families`](https://developers.trawle.com/api/list-families)        |

## Prompt

`daily_job_search` sets up the [daily new-jobs check](https://developers.trawle.com/guides/daily-new-jobs): it resolves the user's places and roles to ids once, runs the search, and keeps `nextSince` so each run only reports jobs that are new.

## Until then

Agents can use the API today: it needs no key, and the [OpenAPI spec](https://developers.trawle.com/openapi.json) and [llms.txt](https://developers.trawle.com/llms.txt) describe it in full.