Jobo for Crush
Live job data without leaving your coding agent.
Jobo's job search server is a remote MCP server, and Crush can connect to it directly. Once it is added, the agent in Crush can search a live index of jobs collected from employer career sites, look up valid filter values, and read full listings — useful when you are building a job board, a sourcing tool or anything else that needs real job data to develop against. You sign in with a Jobo account instead of pasting an API key.
mcp add jobo-job-search --type http --url "https://jobs-mcp.jobo.world/mcp" --oauth true --sessionless trueWhat can you do with the Crush integration?
Give the agent in Crush live job search over MCP, with OAuth sign-in and no API key.
Real listings to build against
Seed a database, fill a fixture file or design a results page with live jobs instead of lorem-ipsum roles. Every result carries its listing URL.
OAuth, not API keys
Crush signs in with your Jobo account. There is no key to copy into a config file, commit by accident, 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.
Filter discovery built in
Lookup tools return the accepted values for every filter with live job counts, so the model uses real values instead of guessing at spellings.
Self-contained job documents
Fetching a job returns responsibilities, qualifications, benefits, compensation and company context as readable prose, so the model can quote it without opening the page.
How do you connect Crush to Jobo?
One-time setup. After this, the data keeps itself current.
- 1
Add the server to Crush
Add this entry to ~/.config/crush/crushrc for every project, or ./.crushrc for one project. If the file already lists other servers, add the jobo-job-search entry alongside them.
bashmcp add jobo-job-search --type http --url "https://jobs-mcp.jobo.world/mcp" --oauth true --sessionless true - 2
Sign in to Jobo
With --oauth true, Crush registers itself with the server and runs the authorization flow. Complete the Jobo sign-in in the browser.
- 3
Check the tools loaded
Start crush and confirm jobo-job-search appears as connected. crush logs shows connection errors if it does not.
- 4
Ask for jobs
“Use Jobo to find 20 remote senior backend roles posted this week and write them to seed.json in my app's job schema.” The agent picks the filters, runs the search and works with real listings instead of invented fixtures.
What goes wrong with Crush, and how do you fix it?
The things people actually hit when running Jobo in Crush.
- Crush keeps asking to authenticate
- The access token expired and was not refreshed. Remove the server and add it again; the server issues a standard OAuth challenge and the sign-in starts over.
- A search returns a credits message instead of jobs
- Searches are billed per job returned, against your plan's included jobs first and then your wallet. When neither covers the request nothing is billed, and the tool returns the amount needed with a top-up link.
- The model invents job listings
- Ask it to include the listing URL for each role. Every real result has one, so a role without a URL did not come from a tool call.
How do you get the most out of Crush?
- crushrc is a shell-style file, not JSON. Add the line as it is.
- --sessionless true tells Crush the server does not issue session IDs, which this one does not.
- Name the filters you care about — location, seniority, work model, salary floor. The model maps them onto real parameters, which is far more accurate than free text.
- Keep result counts small while exploring. Search bills per job returned; reading a single listing is free.
- For production code, call the REST API or an SDK with an API key. The MCP server is for the agent's own lookups while you build.
Which other Jobo integrations work alongside Crush?
Start pulling live jobs into Crush
One Jobo account covers every integration, and the same job index sits behind all of them.













