Get in touch
機械可読For agents

Read Blossom
without scraping.

What this site publishes (company facts, products and the journal) is also served as plain text, Markdown and JSON, from the same source as these pages. Read-only, no sign-in.

同じ内容、別の形。Same content, plainer formats.
— Machine-readable views
入口Start here

Seven
entry points.

/llms.txt
Index of the site and its newest published posts.
/api/public/v1/openapi.json
OpenAPI 3.1 description of the read-only API below.
/api/public/v1/posts
Published posts as JSON, newest first.
/api/public/v1/site
Company facts, offices, products and key pages as JSON.
…/index.md
Markdown view of /, /products, /blog and any /blog/{slug} — append /index.md to the page URL.
/feed.xml
RSS feed of the journal.
/sitemap.xml
Every public page, with real modification dates for posts.
例Examples

Requests that work.

curl "https://blossomai.co/api/public/v1/posts?category=Research&limit=5"
curl "https://blossomai.co/api/public/v1/posts?cursor=<nextCursor from the previous page>"
curl "https://blossomai.co/api/public/v1/posts/discipline-stability-hidden-state"
curl "https://blossomai.co/blog/discipline-stability-hidden-state/index.md"

Lists answer { data, nextCursor, limit }. Pass nextCursor back unchanged for the next page; it is null on the last one.

仕様Behaviour

What to
expect.

  • Only published content. A draft, a post in review or a withdrawn post answers exactly like an unknown slug: 404 not_found.
  • limit is 1–50 (default 20); category matches a category name, case-insensitive. Posts come newest display date first.
  • date is the editorial date shown on the page; modifiedAt is the last change to the published text (null if none); publishedVersion is its revision number.
  • Errors are JSON { error: { code, message } }: 400 invalid_parameter, 400 invalid_cursor, 404 not_found, 503 content_unavailable (retry later).
  • A publish or withdrawal shows up within a minute, caching included: responses carry a max-age that leaves room for the server’s own short cache.
  • No authentication and no rate-limit guarantee. Please keep request rates modest.
利用Use and attribution

Readable is not licensed.

Use, attribution and the policy for AI crawlers are set out in /ai.txt and /robots.txt. Anything beyond that, write to hello@blossomai.co. There is no public write access.

Read the journal →