The client keeps asking to authenticate

The access token expired and the client is not refreshing it. Remove and re-add the connector; the server issues a standard challenge that any compliant client uses to re-authorise.

It cannot be added as a Deep Research connector

Deep Research requires the canonical search and fetch tools. Use the job search server at jobs-mcp.jobo.world — the analytics server at mcp.jobo.world is a separate, SQL-only surface and does not expose them.

The assistant invents job listings

Ask it to cite the listing URL for each role. Every result carries one, and anything without a URL did not come from a tool call.

← All integrationsAI

Let an assistant search the job market directly.

A remote MCP server over Jobo's live job index. Clients sign in with a Jobo account rather than pasting an API key. It exposes the canonical search and fetch pair — the contract ChatGPT Deep Research requires of a connector — alongside a richer filtered search for questions that have structure to them: a location, a salary floor, a seniority, a specific ATS.

Get an API keyMCP registries — coming soon

What it does

Deep Research compatible

Implements the canonical search and fetch tools, so it works as a ChatGPT Deep Research connector rather than only in ordinary chat.

OAuth, not API keys

Clients log in with a Jobo account. Nothing to copy, paste or rotate by hand.

Filters as real parameters

Work model, employment type, experience level, source, skills, industries, salary and date are structured filters, not hints buried in a text query.

Self-contained job documents

Fetching a job returns responsibilities, qualifications, benefits, compensation and company context as readable prose — so an assistant can cite it without opening the page.

Filter discovery built in

A list-filters tool returns every accepted value with live counts, so the model uses real values instead of guessing.

Setting it up

Add the server to your client
In Claude Desktop, add it under Settings → Connectors. In ChatGPT, add it as a connector. Any MCP client supporting Streamable HTTP works.
```
{
  "mcpServers": {
    "jobo-jobs": {
      "url": "https://jobs-mcp.jobo.world/mcp"
    }
  }
}
```
Sign in
The client opens a Jobo login the first time it connects. Approve access and the token refreshes itself from then on.
Ask
“Find senior Rust roles in Europe that sponsor visas and pay over $150k” — the assistant picks the filters, runs the search, and cites the listings it used.

Things that trip people up

The client keeps asking to authenticate: The access token expired and the client is not refreshing it. Remove and re-add the connector; the server issues a standard challenge that any compliant client uses to re-authorise.
It cannot be added as a Deep Research connector: Deep Research requires the canonical search and fetch tools. Use the job search server at jobs-mcp.jobo.world — the analytics server at mcp.jobo.world is a separate, SQL-only surface and does not expose them.
The assistant invents job listings: Ask it to cite the listing URL for each role. Every result carries one, and anything without a URL did not come from a tool call.

Getting the most out of it

Ask for the filters you care about explicitly — the assistant maps them onto real parameters, which is far more accurate than free text.
Ask for listing URLs. It makes fabrication obvious.
For bulk export rather than research, use the Feed API — an assistant is the wrong tool for moving thousands of rows.

Also worth a look

n8n

Search jobs, sync the feed, and trigger workflows when a new role matching your filters is indexed.

Google Sheets

Pull filtered job data straight into a spreadsheet — for market research, comp benchmarking, or a hiring pipeline.