Migrate your docs from Mintlify
Hand your Mintlify repository to a coding agent, review what it changed, keep your URLs working, and deploy a docs site you host yourself.
Guides.One goal at a time.
Migrate to Blume
Migrate your docs from Docusaurus
Move a Docusaurus site's docs to Blume with a coding agent, and check the sidebars, admonitions, and React components it carries over.
Migrate your docs from Fumadocs
Hand your Fumadocs repository to a coding agent, check its meta.ts and component rewrites, keep your docs at /docs, and deploy with every old URL working.
Migrate your docs from Starlight
Move a Starlight site to Blume with a coding agent, keep its pages and URLs, and rebuild the overrides and splash pages that stay Astro work.
Migrate your docs from Nextra
Move a Nextra site's MDX and _meta navigation to Blume with a coding agent, check nested order and React components, and keep every URL working.
Migrate your docs from MkDocs Material
Move an MkDocs Material site's Markdown to Blume, rebuild its nav as folders, rewrite extension syntax, replace its plugins, and check that every old URL still works.
API references
Generate API docs from an OpenAPI spec
Turn an OpenAPI file into a docs site with a page per operation, a request playground, and hand-written guides beside the reference.
Remove internal endpoints from public API docs
Publish a public API reference from the spec your code generates, with an overlay that strips internal operations on every build.
Publish a FastAPI docs site with guides and API reference
Export your FastAPI app's OpenAPI spec without starting a server, and publish it as a separate docs site with a getting-started guide beside the API reference.
Generate documentation from Hono and Zod schemas
A Hono API whose Zod schemas validate requests and generate its OpenAPI spec, published as a docs site with a page per route and a CI check that keeps it current.
Publish NestJS API documentation from Swagger metadata
Turn NestJS controllers and DTOs into a docs site: a Swagger JSON export that needs no database, a page per endpoint with its auth, and a first-request guide beside them.
Publish API documentation from oRPC contracts
Generate an OpenAPI document from your oRPC contract and publish it as an API reference, with a page per operation that updates when the contract changes.
Build an Elysia API documentation website
Describe your Elysia routes and errors so the OpenAPI spec matches the running API, then publish it as a docs site with tutorials beside a page per endpoint.
Generate a Fastify documentation site from route schemas
Turn the JSON Schemas on your Fastify routes into an OpenAPI 3.1 file and a docs site with a page per route, a Try it panel, and guides beside the reference.
Publish Spring Boot API documentation with springdoc-openapi
A Spring Boot API whose Maven build exports and checks its OpenAPI spec, and a static docs site built from that file in CI, with no docs endpoint in production.
Build Django REST Framework docs with drf-spectacular
A DRF API whose drf-spectacular schema generates without warnings, published as an API reference with a first-request tutorial beside it and rebuilt from code in CI.
Generate GraphQL docs from your schema
Turn a GraphQL schema into a docs site with a page per operation and type, linked types and deprecations, and a Try it panel that queries a sandbox.
Document Kafka events with AsyncAPI
Describe your Kafka topics in AsyncAPI, publish event-driven API documentation with a page per operation, and pair it with producer and consumer code you've run.
Write webhook documentation with OpenAPI 3.1
Webhook pages generated from your OpenAPI 3.1 spec, with payloads, signature headers, and acknowledgments, plus a receiving guide with a working Node.js receiver.
Combine multiple OpenAPI specifications in one docs site
An API portal with a reference per service: each OpenAPI spec on its own route and tab, and shared operation names kept apart in URLs, sidebars, and search.
Add SDK examples to generated API documentation
Operation pages that open on your TypeScript and Python SDK calls, with the generated HTTP samples kept or dropped, and a check that catches samples an SDK release broke.
Create interactive API documentation without an OpenAPI spec
Three hand-written MDX endpoint pages with a Try it panel, request samples, and response examples, tested against a local sample API. No spec needed.
Fix CORS errors in an API documentation playground
Reproduce a blocked Try it request, read the preflight the browser sends, and fix it with a CORS policy on your API or Blume's built-in proxy.
Keep API documentation in sync with backend changes in CI
A GitHub Actions workflow that exports your OpenAPI spec from the backend, fails when it's stale, invalid, or breaking, and rebuilds the API reference from it.
Content sources
Turn a Notion database into a docs site
Keep writing in Notion and publish a searchable docs site from a database, with a Status property that decides what goes live.
Publish Obsidian notes as a docs site
Publish a vault of Obsidian notes as a docs site, with wikilinks, heading links, and images resolved at build time.
Use Sanity as a CMS for developer documentation
Editors write guides in Sanity Studio while developers keep Markdown in Git, and one docs site serves both, with drafts only in protected previews.
Build a documentation website with Payload CMS
Publish a Payload collection as a searchable docs site, with Lexical pages, images, and code blocks, draft previews, and a rebuild whenever an editor publishes.
Publish developer documentation from Contentful
Model docs pages and code blocks in Contentful, review drafts with a preview token, and publish a static docs site that rebuilds whenever an editor publishes.
Use Strapi to manage a documentation website
Editors write and publish in Strapi, readers get a static docs site, and every publish starts a fresh build, with drafts kept off production.
Build one docs portal from multiple GitHub repositories
One docs site that pulls each repository's docs folder from GitHub at build time, under its own path, pinned to a release that moves only when you change it.
Combine Markdown and Notion in one documentation site
One docs site where engineers write Markdown in Git and teammates edit Notion, with a section each, checks that keep their URLs apart, and one build for both.
Turn GitHub releases into a product changelog
Write release notes once on GitHub and get a changelog timeline, a page per release, and an RSS feed on your docs site, rebuilt every time you publish.
Hosting
Deploy Markdown docs to GitHub Pages
Turn the Markdown in your repository into a searchable docs site on GitHub Pages that rebuilds every time you push.
Serve a separate docs site under /docs in Next.js
Deploy Blume as its own static site under /docs, proxy it through one Next.js rewrite, and check pages, assets, canonical URLs, and llms.txt through your domain.
Self-host documentation with Docker and Node.js
A Docker image that serves your docs and MCP endpoint from Node.js, run with Compose behind Caddy, with a health check and tests for pages and discovery files.
Deploy docs and an MCP server to Cloudflare Workers
One Cloudflare Worker that serves your docs as static assets, answers Markdown requests at the same URLs, and hosts an MCP endpoint coding agents connect to.
Protect private docs with Cloudflare Access
A private docs site on Cloudflare Workers that only your company can open, with every page, Markdown copy, search index, and agent endpoint checked behind Access.
Add documentation to a monorepo without breaking app builds
A docs workspace beside your app that shows live examples from a shared package, builds with Turborepo, and deploys on its own, skipping commits it isn't part of.
Host static documentation on S3 and CloudFront
A private S3 bucket behind CloudFront that serves every page at its clean URL, Markdown and JSON with the right types, real 404s, and a cache each deploy clears.
Choose static or server-rendered documentation
Map each docs feature to static files or a live endpoint, build the same site both ways, and check what each build ships and what you would have to run.
AI and agents
Add an MCP server to your docs
Let coding agents search and read your docs from inside the editor, through an MCP server that deploys with your site.
Rate-limit a public documentation AI assistant
A per-reader limit on your docs assistant that holds across serverless instances, a load test that shows the 429s and the reset, and a bot check for scripts.
Add an AI assistant that answers from your documentation
An in-page assistant that answers from your docs and links its sources, a support handoff for what they don't cover, and an evaluation set that shows where it fails.
Embed your documentation assistant inside your product
A chat panel in your React app that streams answers from your Blume docs, knows which page covers the current screen, and links every source back to the docs.
Run a documentation assistant with Ollama
A docs assistant that answers from your pages with a model running in Ollama beside your docs server, sized for your hardware and measured against a hosted model.
Test whether your documentation can answer user questions
An evals file built from real support questions, a failing run traced to the page that should answer, the fix that turns it green, and a CI gate on docs changes.
Keep documentation updated from merged code changes
A coding agent reads your merged pull requests, fixes only the docs they made wrong, shows its evidence in a pull request for review, and runs weekly in GitHub Actions.
Publish a tested agent skill for your SDK
A SKILL.md grounded in your docs that walks coding agents through your SDK's setup, published with your docs site and tested in a fresh project against expected outcomes.
Index documentation for RAG without scraping HTML
A sync script that indexes your docs from their JSON API, embeds only changed pages, and drops removed ones, plus a query that cites the exact section each answer came from.
Serve HTML and Markdown from the same documentation URL
Browsers get HTML and agents that send Accept: text/markdown get the page's Markdown at the same URL, checked by a script that tests headers, content, and caching.
Search
Add Pagefind search to a static documentation site
Switch your docs search to Pagefind, build the index with the site, and check headings, languages, long pages, and exclusions with a query script you rerun on every build.
Self-host documentation search with Typesense
A Typesense server you run, scoped keys for your build and your readers, a docs collection rebuilt on every production build, and synonyms relinked after each sync.
Find missing documentation from searches with no results
A saved report of the searches your docs can't answer, sorted into pages to write and words to add, with a keyword fix checked before and after.
Choose local, hosted or semantic search for your docs
Run one fixed query set through local, hosted, and semantic search, compare what each finds and what it takes to run, and keep the engine your docs need.
Make docs search work in Chinese, Japanese and Korean
A four-language test corpus, a query list to check it against, and a search setup that finds Chinese, Japanese and Korean words, with results for each engine.
SEO and discoverability
Check whether AI search crawlers can access your docs
A repeatable check of robots.txt, CDN bot rules, status codes, and indexing signals for the crawlers behind ChatGPT search and Google's AI features.
Measure AI search referrals and citations for your docs
Track AI citations, assistant referrals, and the reader actions that follow as separate numbers, with a dated baseline you compare every month.
Add a technical SEO check to your documentation build
A test project with planted SEO failures, an audit that points each one at its source line, the fixes, and a GitHub Actions job that blocks a bad deploy.
Find out why Google isn't indexing your docs pages
Read the reason Search Console gives, trace robots.txt, noindex, and canonical causes to the file behind them, and confirm each fix on the live site.
Move documentation URLs without breaking old links
Map every old URL to its new page, serve real HTTP redirects from your host, check each one on the live site, and keep a way back if the move goes wrong.
Fix hreflang and canonical URLs on multilingual docs
A two-language docs site where every translation is canonical to itself and linked by hreflang, untranslated pages defer to the original, and an audit proves it.
Writing and maintenance
Version your docs for a breaking release
Snapshot your docs before a breaking release, so readers on the old and new versions each find the right instructions.
Translate your docs and catch stale translations
Translate your docs with a coding agent, then keep them current with a CI check that fails when a source page changes.
Reuse Markdown snippets and variables across documentation
Three guides that share one setup section through MDX includes, with one intentional variation, version strings defined once, and a way to check every page a change reaches.
Build an engineering handbook with RFCs, ADRs, and runbooks
A handbook of RFCs, decision records, and runbooks where the build rejects a missing owner or status, and agents can list documents by type, team, or status.
Catch broken Markdown links in GitHub Actions
A pull request check that fails on broken docs links and heading anchors and annotates each one on its line, plus a weekly job that probes external URLs.
Add live React component demos to your MDX docs
A live component demo beside the code that uses it, with a props table, a text version for readers and agents that can't run it, and a keyboard test on the production build.
Build a developer portal for multiple products
A two-product developer portal where each product has its own tab, sidebar, quickstart, and API reference, shared pages stay one click away, and search results name the product.
Collect documentation feedback and turn it into fixes
A page-level helpfulness report in PostHog, the reader comments behind your weakest pages, and a pull request that records why each fix was made.
Offer PDF and EPUB downloads of your docs pages
Readers can save any docs page as a PDF or an EPUB, and you know what each format keeps, from long code blocks and tabs to images and diagrams.
Create architecture documentation with Mermaid diagrams
An architecture page with a component flowchart and a request sequence diagram, kept as Mermaid source in Git, plus text that explains the system wherever diagrams don't render.
Upgrade your docs with Blume.
Install today and ship a production-grade docs site in minutes. Free and open source, forever.
npx blume init