for developers

The jobs API

Every live data science, ML engineering and AI research role on this board, as JSON. Free, no contract, no sales call. Use it in your app, your newsletter, your Discord bot, your own job board or your AI assistant — the only thing we ask for is a link back.

Quick start

No key needed to try it:

curl "https://api.finddatasciencejobs.com/v1/jobs?limit=10"

You get a JSON object with a jobs array, your remaining quota, and the attribution terms. Newest roles first, open listings only — closed ones drop out on their own.

Limits

Tier	Per day	How
Anonymous	50	Just call it.
Free key	500	One request, below. No card, no confirmation email.
More than that	Talk to us	If you are building something real, we would rather say yes.

The count is in jobs, not requests — asking for ten jobs once and asking for one job ten times cost exactly the same. A filter that matches nothing costs nothing, and so does a 404.

One count covers both the JSON feed and the MCP server below. They are two ways to read the same board, not two allowances.

Get a key

curl -X POST "https://api.finddatasciencejobs.com/v1/keys" \
  -H "Content-Type: application/json" \
  -d '{"email":"you@example.com","site":"https://your-site.com"}'

The key comes back in the response and is not shown again — save it. Send it on every call:

curl "https://api.finddatasciencejobs.com/v1/jobs?limit=100&since=2026-09-01" \
  -H "X-API-Key: fdsj_your_key_here"

Filters

Parameter	What it does
`limit`	How many jobs to return, up to 100 (and up to whatever you have left today).
`since`	`YYYY-MM-DD` — only roles posted on or after that date. The polite way to poll daily.
`search`	Free text across title, company and description.
`remote`	`true` / `false`.
`country`	Country slug, e.g. `usa`, `india`, `united-kingdom`.
`role_type`	e.g. `data_scientist`, `ml_engineer`, `data_engineer`.
`seniority`	e.g. `junior`, `mid`, `senior`.

There is no page parameter on purpose. Use since to take only what is new since your last run — it is cheaper for you and kinder to us than walking the whole board every night.

A single job

curl "https://api.finddatasciencejobs.com/v1/jobs/data-scientist-stripe-7"

Takes the slug from a job’s url. Costs one job from your daily count.

Use it from an AI assistant (MCP)

The board is also an MCP server, so an assistant can search it for itself rather than being told where the docs are. Nothing to install and nothing to sign in to — add this URL as a connector:

https://api.finddatasciencejobs.com/mcp

Tool	What it does
`search_jobs`	The filters above, returning a summary of each match. Costs one job per result.
`get_job`	One role in full, description included. Costs one.
`list_filters`	The country slugs and role types the board actually uses, so nothing has to guess. Free.

Using a key raises the daily count exactly as it does on the feed. Send it as the X-API-Key header, or — since many clients only let you paste a URL — put it on the end of that one:

https://api.finddatasciencejobs.com/mcp?key=fdsj_your_key_here

The licence does not change because an assistant is the one reading: if it shows someone a role, it should link to that role’s page here. Every tool answer carries the terms and the link with it.

The terms

Use the data however you like — commercially, in a paid product, in a competing job board. We mean it. Three conditions, and they are all the same condition:

Link back. Each job comes with a url on this site. If you show the job, link to that URL as an ordinary followable link — no rel="nofollow".
Name us as the source. Once per page is plenty. There is ready-made HTML in every response under attribution.
Send applicants to the job’s page here. The apply_url we give you is that page. Don’t swap it for an employer link you found another way.

That is the whole deal: you get clean, enriched, deduplicated listings for nothing, and we get the visit back. Keys that strip the attribution get revoked — it is the only thing we police.

What you get per job

Title, company (with website and logo), location, country and cities, remote flag, employment type, seniority, role type, skill tags, salary range where the employer published one, a short AI-written summary, the full description, and the date it was posted.

Two things we don’t hand over: which site a listing came from, and any employer contact address. The first is ours; the second isn’t ours to give.

Fair use

The feed is one machine-readable endpoint so that nobody has to write a scraper. If your key is not enough for what you are building, get in touch rather than working around it — the answer is usually yes.