Reed MCP server: let your AI agent search UK jobs with filters

#What your agent gets

The tool takes a required keywords input, for example python, data scientist or warehouse. The optional inputs are location, distance_miles, min_salary, max_salary, contract, hours, include_ads and page.

#Inputs and output

The result has keywords, location, total, page and jobs. Each job has title, company, location, contract, hours, working, salary_from, salary_to, salary_per, posted_at, expires_at, easy_apply, promoted and url. salary_per shows the unit of the salary: per year, per day or per hour. total counts all matches and can be null.

#How a request flows

1. You ask your agent: "Find permanent Python jobs in London, full time, with salary and closing date."
2. The agent calls find_recipe, then run with reed.co.uk/search-jobs, keywords set to python, location set to london, contract set to permanent and hours set to full_time.
3. wield-mcp passes the run through the local bridge to the Wield extension.
4. The extension sends the search request from your browser, so the recipe runs in your own browser and not on a server.
5. The jobs go back to your agent. Nothing passes through wield.run.

#What we saw in our check

On 7 November 2026, a permanent search for python in London passed all 6 checks in one request, in 408 ms, and returned 1,064 matches. We also checked a contract search on page 2, an hourly part-time search and a search with no results.

The recipe reads the search results that the page publishes in its data. A promoted job in the list is dropped unless you set include_ads to true.

#Salary units

Reed lists salaries per year, per day or per hour. The recipe reads the unit and returns it in salary_per, so a value such as 650 to 750 per day is not mistaken for an annual figure. Our check read an hourly rate of £14.70 per hour the same way.

#What leaves your computer

The recipe makes one request, to www.reed.co.uk, for one results page, as the site serves it to a visitor. The request leaves from your browser. Nothing passes through wield.run. Wield logs each request on your computer only, with method, host, path, status and timing, and never logs bodies or query values. Telemetry is off by default. Wield never reads, copies or sends your cookies or passwords.

#Reed's terms, in plain words

Reed's robots.txt allows the search pages. It disallows /api/ and search URLs with sortBy=, so the recipe never sorts. It has no rules for Claude agents or AI content signals. The recipe makes one request per call. It does not submit applications, set up job alerts or save jobs, and it makes no purchases. Our review of Reed's terms is not finished, and this is not legal advice. Use the recipe for one search that you start yourself, in your own browser, and not for bulk collection of listings. Read the site's terms yourself before you use this recipe in any other way.

#Limits and what is not working yet

• Read only. The recipe searches. It does not apply, set up job alerts or save jobs.
• No sorting. The recipe never sorts results.
• Distance needs a location. distance_miles has no effect without location.
• Promoted jobs. Dropped by default. Marked when include_ads is true.
• Run limits apply. A run has at most 12 steps, 40 requests, 5 MB per response and 60 seconds.

#FAQ

#Can the agent apply for a job?

No. The recipe only searches. Open the job's url to apply on Reed yourself.

#Does it need a Reed login?

No. The search works without an account.

#What does distance_miles do?

It limits results to a radius around location, in miles. It applies only when you also set a location.

#Does it show promoted jobs?

Not by default. Set include_ads to true to keep them. Each kept job is marked as promoted.

#Which AI apps work with it?

Any MCP client. npx -y wield-mcp init sets up Claude Desktop, Claude Code, Cursor and Codex. Other MCP clients work with the standard stdio config.

#Where do the results go?

To your agent, on your own computer. The results never pass through wield.run.