usage

Adding Documentation

Fetch and index documentation from URLs, GitBook and Mintlify sites, GitHub repos, and local files into a searchable local Library.

Updated

One ingest command

docmancer docs add handles URLs, GitHub repositories, and local paths. It writes into the same default local hybrid index Docmancer uses for memory: SQLite FTS5 plus dense vectors from the vendored model in sqlite-vec. Sparse retrieval is available only on the scale retrieval profile, which uses Qdrant and SPLADE.

Add from a URL

docmancer docs add https://docs.example.com

Docmancer auto-detects the docs platform and chooses the best fetching strategy:

PlatformDetectionStrategy
GitBookllms-full.txt endpointFull-text download
Mintlifyllms.txt or sitemap.xmlSitemap crawl
GitHubRepository URLREADME + docs directory extraction
Generic websitemap.xml or nav crawlPage-by-page fetch

Force a specific provider:

docmancer docs add https://docs.example.com --provider mintlify

Add from a GitHub repo

docmancer docs add https://github.com/owner/repo

Extracts the README and any docs/ directory content.

Ingest local files

docmancer docs add ./my-internal-docs

Supported file formats: Markdown (.md, .mdx), plain text (.txt), PDF (.pdf), DOCX (.docx), RTF (.rtf), and HTML (.html, .htm). Every parser ships in the base install, so none of these needs an extra.

Options

docmancer docs add flags:

FlagDefaultDescription
--providerautoauto, gitbook, mintlify, web, github, crawl4ai
--strategyautoForce discovery strategy (e.g. llms-full.txt, sitemap.xml, nav-crawl)
--max-pages500Limit pages fetched (web provider)
--browseroffUse Playwright for JS-heavy sites (needs docmancer[browser])
--fetch-workersautoNumber of concurrent page fetch workers
--include / --excludeunsetGlob patterns for a local path
--formatallRestrict local ingest to specific formats
--recursive / --no-recursiverecursiveRecurse through local directories
--skip-knownoffSkip local files whose content hash is already indexed
--no-vectorsoffIndex local files with FTS5 only
--recreateoffDrop and rebuild the index for this source

List and update existing sources

docmancer docs list
docmancer docs sync
docmancer docs sync https://docs.example.com

docs list shows every indexed source (add --all for individual pages and files). docs sync with no argument refreshes every source; pass a specific URL or path to update only that one. Updates reuse the content-hash-keyed embeddings cache, so unchanged sections skip re-embedding.

Remove a source when you no longer need it:

docmancer docs remove https://docs.example.com
docmancer docs remove --all