# Ashby

Source: https://career-ops.org/docs/reference/portals/ashby (canonical HTML, identical content)

> Ashby is the modern ATS choice for Series A–C startups in the AI and dev-tools space. career-ops hits the public job-board API directly.



## What it does [#what-it-does]

The ashby provider in providers/ashby.mjs makes HTTP requests to api.ashbyhq.com/posting-api/job-board/`{slug}` and parses the resulting JSON. The slug is the company's Ashby job board identifier (e.g. 'anthropic' for Anthropic).

## When to use it [#when-to-use-it]

Use Ashby scanning when targeting AI labs, infrastructure companies, and modern startups. Anthropic, Vercel, Replicate, and many YC companies use Ashby. About 40+ pre-configured slugs ship in templates/portals.example.yml.

## Example [#example]

```
In your portals.yml: `- provider: ashby\n  slug: anthropic\n  careers_url: https://jobs.ashbyhq.com/anthropic` — then scan picks up active Anthropic roles.
```

## What the scanner does [#what-the-scanner-does]

The scanner reads your enabled Ashby slugs from portals.yml, calls Ashby's public job-board API for each one, and merges the fresh roles into your pipeline with no browser, no login, and zero LLM tokens spent on the fetch itself. A single `node scan.mjs` run covers your Ashby, Greenhouse, and Lever boards together, then filters the results by the title keywords you configured.

```bash title="Terminal"
node scan.mjs
```

```txt title="Output"
Portal Scan - 2026-07-06
Offers found: 41 total
Filtered by title: 6 relevant
Duplicates: 3 (already evaluated or in pipeline)
New added to pipeline.md: 3

  + Anthropic | Applied AI Engineer | ashby
  + Replicate | Inference Engineer  | ashby
```

## How career-ops talks to Ashby [#how-career-ops-talks-to-ashby]

career-ops queries Ashby's `ApiJobBoardWithTeams` GraphQL endpoint with your board slug and reads the `jobPostings` list straight from the JSON response. For each posting it pulls the title, an id used to build the public apply URL, the location name, the employment type, and a compensation summary when the board exposes one. Because this is a structured API call rather than page scraping, the scan is fast, stays inside rate limits, and spends no model tokens. Tokens are only spent later, during the evaluate step, when you decide a role is worth a closer look. Re-running the scan is safe: every URL is deduplicated against your scan history, your evaluated applications, and the current pipeline, so a role you have already seen is never added twice.

## Gotchas [#gotchas]

Ashby's job board API returns full JD text directly, no follow-up request needed. Evaluation is slightly faster than for Greenhouse listings.

## Related [#related]

* [greenhouse](https://career-ops.org/docs/reference/portals/greenhouse.md)
* [lever](https://career-ops.org/docs/reference/portals/lever.md)