Support HTTP Basic client authentication in OAuth password flow
Some OAuth servers require client credentials in an HTTP Basic Authorization header instead of the request body. The password flow dialog now offers a Client Authentication select to choose between the two methods, as described in RFC 6749, section 2.3.1.
Generated pages are grouped in a directory named after the root module. Set groupBy: 'none' to place them at the root of the source instead, the root module becomes index.mdx:
Pages are compiled in memory with page.data.load(), render() includes the default MDX components of Fumadocs UI and fumadocs-python/components. Links between modules and classes are resolved from the loader the source is attached to, and loaderPlugin() badges generated pages in the page tree.
convert() shares the same page builder. write() takes the output directory directly, and keeps the package name in file paths:
Read structured data from page.data.structuredData()
Search indexing no longer falls back to (await page.data.load()).structuredData. Runtime content sources expose structuredData() on page data instead, sharing the compile with load():
The renderer returned by load() still carries structuredData, existing code keeps working.
@fumadocs/base-ui@16.15.1
fumadocs-core@16.15.1
fumadocs-ui@16.15.1
Forward dynamic loader from fumadocs-core/source
Read structured data from page.data.structuredData()
Search indexing no longer falls back to (await page.data.load()).structuredData. Runtime content sources expose structuredData() on page data instead, sharing the compile with load():
The renderer returned by load() still carries structuredData, existing code keeps working.
Read structured data from page.data.structuredData()
Search indexing no longer falls back to (await page.data.load()).structuredData. Runtime content sources expose structuredData() on page data instead, sharing the compile with load():
The renderer returned by load() still carries structuredData, existing code keeps working.
Read structured data from page.data.structuredData()
Search indexing no longer falls back to (await page.data.load()).structuredData. Runtime content sources expose structuredData() on page data instead, sharing the compile with load():
The renderer returned by load() still carries structuredData, existing code keeps working.
Read structured data from page.data.structuredData()
Search indexing no longer falls back to (await page.data.load()).structuredData. Runtime content sources expose structuredData() on page data instead, sharing the compile with load():
The renderer returned by load() still carries structuredData, existing code keeps working.
Redesign source API
Content sources can hook into the static loader they are attached to, and dynamic sources can opt out of the loader's in-memory file cache.
configureStatic runs when a source is attached to loader(), and again whenever dynamicLoader() builds a new static loader:
export function createMySource(): DynamicSource { return { cache: 'custom', async files() { return loadFiles(); }, configureStatic({ loader, source }) { // `loader` is the created static loader // `source` is the record key when using named sources }, configure(loader, { source }) { loader.invalidate(); }, };}
cache: 'memory' (default): files() is called once until invalidate().
cache: 'custom': the source caches itself. dynamicLoader() re-runs files() on get() and rebuilds only when the file list is shallowly different (by identity).
Integrations
GraphQL cross-links are generated from the attached loader instead of a baseUrl option on staticSource(). Local, OpenAPI, and AsyncAPI dynamicSource() use cache: 'custom' and reuse generated files by identity until invalidate().
Sanity now uses cache: 'custom' when given a sanityFetch from next-sanity/live, calling invalidate() in draft mode is no longer needed.
Scope lastModified git log to the content directory
git log is scoped to the collection's content directory instead of buffering the repository's entire history in every worker.
Fix Vite dev server crash on declaration-only dependencies
The injected Vite config no longer pre-bundles packages without runtime JavaScript, such as @types/mdx. Pre-bundling them made esbuild parse .d.ts files and fail on imports that only exist in type space, crashing the dev server.
Encode import.meta.glob query values
The Vite codegen passed the query to import.meta.glob as an object, letting the bundler serialize it. Rolldown inlines the values as-is, so a macro id such as src/lib/source.ts#docs left an unescaped / in the content file's module id and relative imports from that module (e.g. images from ) failed to resolve, since the importer's directory is derived from the raw id.
The query is now serialized (and percent-encoded) by Fumadocs itself, matching what the Node.js codegen already did.
Redesign source API
Content sources can hook into the static loader they are attached to, and dynamic sources can opt out of the loader's in-memory file cache.
configureStatic runs when a source is attached to loader(), and again whenever dynamicLoader() builds a new static loader:
export function createMySource(): DynamicSource { return { cache: 'custom', async files() { return loadFiles(); }, configureStatic({ loader, source }) { // `loader` is the created static loader // `source` is the record key when using named sources }, configure(loader, { source }) { loader.invalidate(); }, };}
cache: 'memory' (default): files() is called once until invalidate().
cache: 'custom': the source caches itself. dynamicLoader() re-runs files() on get() and rebuilds only when the file list is shallowly different (by identity).
Integrations
GraphQL cross-links are generated from the attached loader instead of a baseUrl option on staticSource(). Local, OpenAPI, and AsyncAPI dynamicSource() use cache: 'custom' and reuse generated files by identity until invalidate().
Sanity now uses cache: 'custom' when given a sanityFetch from next-sanity/live, calling invalidate() in draft mode is no longer needed.
Remember verified file URLs in the Notion file handler
Every asset request cost between 2 and 34 Notion API calls, so a page with a handful of images tripped the rate limit with two concurrent visitors. The handler now remembers a verified signed URL per block until shortly before Notion's expiry_time.
Do not cache rejected promises
The dynamic loader's files(), Notion's page load(), createFromSource's index build, and Shiki factory init retry on the next call after a transient failure, instead of returning the same rejection forever.
Redesign source API
Content sources can hook into the static loader they are attached to, and dynamic sources can opt out of the loader's in-memory file cache.
configureStatic runs when a source is attached to loader(), and again whenever dynamicLoader() builds a new static loader:
export function createMySource(): DynamicSource { return { cache: 'custom', async files() { return loadFiles(); }, configureStatic({ loader, source }) { // `loader` is the created static loader // `source` is the record key when using named sources }, configure(loader, { source }) { loader.invalidate(); }, };}
cache: 'memory' (default): files() is called once until invalidate().
cache: 'custom': the source caches itself. dynamicLoader() re-runs files() on get() and rebuilds only when the file list is shallowly different (by identity).
Integrations
GraphQL cross-links are generated from the attached loader instead of a baseUrl option on staticSource(). Local, OpenAPI, and AsyncAPI dynamicSource() use cache: 'custom' and reuse generated files by identity until invalidate().
Sanity now uses cache: 'custom' when given a sanityFetch from next-sanity/live, calling invalidate() in draft mode is no longer needed.
Read files in bounded chunks during a cold scan
getFiles() awaits each chunk before starting the next, instead of starting the entire tree concurrently.
@fumadocs/base-ui@16.15.0
fumadocs-core@16.15.0
fumadocs-ui@16.15.0
Redesign source API
Content sources can hook into the static loader they are attached to, and dynamic sources can opt out of the loader's in-memory file cache.
configureStatic runs when a source is attached to loader(), and again whenever dynamicLoader() builds a new static loader:
export function createMySource(): DynamicSource { return { cache: 'custom', async files() { return loadFiles(); }, configureStatic({ loader, source }) { // `loader` is the created static loader // `source` is the record key when using named sources }, configure(loader, { source }) { loader.invalidate(); }, };}
cache: 'memory' (default): files() is called once until invalidate().
cache: 'custom': the source caches itself. dynamicLoader() re-runs files() on get() and rebuilds only when the file list is shallowly different (by identity).
Integrations
GraphQL cross-links are generated from the attached loader instead of a baseUrl option on staticSource(). Local, OpenAPI, and AsyncAPI dynamicSource() use cache: 'custom' and reuse generated files by identity until invalidate().
Sanity now uses cache: 'custom' when given a sanityFetch from next-sanity/live, calling invalidate() in draft mode is no longer needed.
Return heading and text results from the Algolia client
algoliaClient grouped hits into page, heading and text results, then dropped everything except pages. All three are now returned with highlighting, matching the other search clients.
Fix locale-only pages leaking into other locales
i18n storages no longer share folder arrays with the fallback locale. Locale-only pages previously appeared in every locale's page tree as duplicate nodes.
Index pages by URL
getPageByHref resolves absolute URLs through an index instead of scanning all pages on every call.
Do not cache rejected promises
The dynamic loader's files(), Notion's page load(), createFromSource's index build, and Shiki factory init retry on the next call after a transient failure, instead of returning the same rejection forever.
Highlight only the search results within limit
limit bounds the returned hits, so search now stops at that many results instead of highlighting every matched page and slicing afterwards.