Skip to content
Agentmart
Submit a tool

DocPull

Sync public web sources into cited context packs for AI agents, RAG, and MCP clients.

Connector by raintree-technology · GitHub account verified by the MCP Registry

IndexedMITLocal (stdio)★ 29
Write a review

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/raintree-technology-docpull

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

Sync public web sources into cited context packs for AI agents, RAG, and MCP clients.

Use cases: knowledge-base, web-scraping, documentation

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list, install config and code. Checked Oct 25, 2026.

Tools
18 read · 19 write · 1 delete
Secrets
None asked for
Install scripts
None found
Runs as
Python package

Tools by what they do 38

Read from the source code; the running server may list different tools.

Read 18 · Look things up without changing anything.

  • context_query
  • explain_routes
  • fetch_url
  • graph_neighbors
  • graph_query
  • graph_status
  • grep_docs
  • list_indexed
  • list_sources
  • pack_brief
  • pack_citations
  • pack_diff
Show 6 more tools
  • pack_entities
  • pack_score
  • pack_search
  • read_doc
  • serve_pack_status
  • validate_policy

Write 19 · Create, change, send or run things.

  • add_source
  • audit_pack
  • brand_pack
  • ensure_docs
  • export_pack
  • graph_build
  • graph_refresh
  • image_pack
  • intelligence_bundle
  • pack_prepare
  • policy_pack
  • product_pack
Show 7 more tools
  • refresh_pack
  • relationship_pack
  • render_url
  • screenshot_pack
  • styleguide_pack
  • website_pack
  • workflow_run

Delete 1 · Delete, drop, revoke or wipe things.

  • remove_source

Network hosts it talks to 60

  • 0f21cc82.preview.developers.cloudflare.com · found in code
  • 9df0cba6.preview.developers.cloudflare.com · found in code
  • api.exa.ai · found in code
  • api.firecrawl.dev · found in code
  • api.parallel.ai · found in code
  • api.tavily.com · found in code
  • arxiv.org · found in code
  • cheatsheetseries.owasp.org · found in code
  • chromereleases.googleblog.com · found in code
  • cisa.gov · found in code
  • community.vercel.com · found in code
  • csrc.nist.gov · found in code
  • curl.se · found in code
  • developer.hashicorp.com · found in code
  • developers.cloudflare.com · found in code
  • discuss.kubernetes.io · found in code
  • docker.com · found in code
  • docpull.dev · found in code
  • docs.anthropic.com · found in code
  • docs.astral.sh · found in code
  • docs.djangoproject.com · found in code
  • docs.docker.com · found in code
  • docs.example.test · found in code
  • docs.firecrawl.dev · found in code
  • docs.github.com · found in code
  • docs.openssl.org · found in code
  • docs.parallel.ai · found in code
  • docs.stripe.com · found in code
  • docs.tavily.com · found in code
  • exa.ai · found in code
Show 30 more hosts
  • fastapi.tiangolo.com · found in code
  • firecrawl.dev · found in code
  • git-scm.com · found in code
  • git.postgresql.org · found in code
  • github.blog · found in code
  • github.com · found in code
  • kubernetes.io · found in code
  • mta.openssl.org · found in code
  • nextjs.org · found in code
  • nginx.org · found in code
  • numpy.org · found in code
  • nvd.nist.gov · found in code
  • openssl.org · found in code
  • postgresql.org · found in code
  • raintree-technology.github.io · found in code
  • raw.githubusercontent.com · found in code
  • schemas.openxmlformats.org · found in code
  • services.nvd.nist.gov · found in code
  • spec.openapis.org · found in code
  • sqlite.org · found in code
  • tc39.es · found in code
  • url.spec.whatwg.org · found in code
  • usenix.org · found in code
  • v1-27.docs.kubernetes.io · found in code
  • v1-28.docs.kubernetes.io · found in code
  • v1-29.docs.kubernetes.io · found in code
  • v1-32.docs.kubernetes.io · found in code
  • v1-33.docs.kubernetes.io · found in code
  • vercel.com · found in code
  • wiki.openssl.org · found in code

Files it touches

  • Your home directory (~)
  • ~/.config/docpull-mcp/sources.yaml
  • Any path the agent passes to its tools

Tools (38)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

  • add_sourceAdd or update a user source

    Add or update a user source alias in the writable sources.yaml. Refuses to shadow a builtin alias unless force=true. URL is HTTPS-only and validated against the same SSRF rules as fetch_url. Use list_sources to confirm the change.

  • audit_packAudit a local pack

    Write pack.audit.json and PACK_AUDIT.md with deterministic quality dimensions.

  • brand_packBuild a brand pack

    Build an evidence-backed brand pack through the common workflow protocol.

  • context_queryQuery locked contextread-only

    Query verified project lock archives offline. Return deduplicated passages and exact citations. max_tokens bounds the canonical JSON response including citations using a conservative UTF-8 byte count; MCP framing is excluded. versions filters explicit metadata.version values; pack_versions filters exact locked run IDs.

  • ensure_docsFetch a source alias

    Fetch Markdown for a named source alias (e.g. 'react', 'nextjs', or a user-added website). Uses a 7-day cache; pass force=true to refresh. Optional profile selects fetch behavior: rag (default, balanced for retrieval), mirror (full archive), quick (fast/shallow), llm (NDJSON chunks). Use list_sources to discover aliases first.

  • explain_routesExplain routesread-only

    Cost: local, $0. Return the cost-ordered acquisition route ladder, budget semantics (budget=0 blocks paid-capable routes before any spend), which provider API keys are set (names/booleans only), and per-route unlock hints. Optional 'url' also lists the local discovery sources that would be attempted. Call this before any paid-capable work to reason about the escalation ladder. Purely local; performs no network probing.

  • export_packExport pack

    Export a local pack to agent-safe JSONL or skill/rule formats.

  • fetch_urlFetch one HTTPS pageread-only

    Fetch a single HTTPS URL and return clean Markdown. No discovery or crawl — the agent-friendly fast path. Returns the page's Markdown with source and detected framework in the header. Optionally chunk the output with max_tokens. Rejects non-HTTPS URLs, localhost, and private IPs. For whole source aliases use ensure_docs instead.

  • graph_buildBuild pack source graph

    Build local cited source graph sidecars for a DocPull pack: graph.json, graph.nodes.ndjson, graph.edges.ndjson, and GRAPH.md.

  • graph_neighborsFind graph neighborsread-only

    List cited neighboring nodes for matching graph entity nodes.

  • graph_queryQuery source graphread-only

    Search graph nodes and cited graph edge evidence without generating an answer.

  • graph_refreshRefresh source graph

    Rebuild local source graph sidecars from the current pack and write graph.diff.json.

  • graph_statusCheck graph freshnessread-only

    Report whether local graph artifacts are missing, current, or stale.

  • grep_docsRegex-search cached Markdownread-only

    Regex search through fetched web-source Markdown. Results are ranked by match density (most matches per file first) and rendered with lines of surrounding context. Each result returns the source alias as the historical library field plus a path relative to that source, so you can feed both fields straight into read_doc. Use ensure_docs first.

  • image_packBuild an image pack

    Build a bounded visual-asset manifest through the common workflow protocol.

Show 23 more tools
  • intelligence_bundleGenerate an intelligence bundle

    Generate a deterministic intelligence.bundle.v1 tracker import with source snapshots, document versions, precise observations, and change candidates.

  • list_indexedList locally cached sourcesread-only

    List web sources that have been fetched to the local Markdown directory, with last-fetched age. Sorted alphabetically.

  • list_sourcesList configured source aliasesread-only

    List configured source aliases, optionally filtered by category. Use this to discover what ensure_docs can fetch.

  • pack_briefBuild pack research briefread-only

    Generate a cited local research brief from a docpull context pack, including source load order, key excerpts, and structured signals.

  • pack_citationsBuild pack citationsread-only

    Build a stable citation/source map for a docpull context pack.

  • pack_diffDiff context packsread-only

    Diff two docpull context packs by URL and content hashes without shelling out.

  • pack_entitiesExtract pack entitiesread-only

    Extract cited entities and structured signals from a docpull context pack locally.

  • pack_preparePrepare pack artifacts

    Write the standard local pack intelligence bundle: pack score, source scores, citations, entities, local search results, and a cited research brief.

  • pack_scoreScore a context packread-only

    Score a docpull context pack for agent-readiness without shelling out.

  • pack_searchSearch a context packread-only

    Search a docpull context pack locally and return cited excerpts.

  • policy_packBuild a policy evidence pack

    Discover policy documents and emit stable clause evidence and textual change candidates; does not provide legal conclusions.

  • product_packBuild a product pack

    Build product and pricing evidence through the common workflow protocol.

  • read_docRead a cached Markdown fileread-only

    Read a Markdown file from a fetched source alias, optionally sliced by line range. The natural follow-up to grep_docs: pass each result's library/source alias and path (path is already relative to the source root) to pull more surrounding context.

  • refresh_packRefresh a local pack

    Refresh the URLs from an existing local pack, write a refreshed snapshot plus refresh.report.json/md, and return a structured diff. Use dry_run=true to plan without network calls.

  • relationship_packBuild a relationship evidence pack

    Extract cited owned-by, operated-by, acquired-by, franchised-by, and invested-in observations. Outputs review candidates and never infers independence from missing evidence.

  • remove_sourceRemove a user sourcecan change or delete data

    Remove a user source alias. Optionally delete its cached Markdown cache (delete_cache=true). Cannot remove a builtin source — to stop using one, just don't call ensure_docs on it.

  • render_urlRender one URL

    Render one public URL through an explicit optional render runtime and write rendered HTML plus rendered_pages.ndjson. Runtimes may require agent-browser, Vercel Sandbox auth, or E2B_API_KEY; no captcha solving or stealth behavior.

  • screenshot_packCapture a screenshot pack

    Capture a screenshot pack through the explicitly trusted local browser gate and common workflow protocol.

  • serve_pack_statusPack server statusread-only

    Return the same health payload the local pack server exposes at /health, without starting a listener.

  • styleguide_packBuild a styleguide pack

    Build styleguide and design-token evidence through the common workflow protocol.

  • validate_policyValidate policyread-only

    Validate a local DocPull policy file and return its non-secret source_policy payload.

  • website_packBuild a website snapshot

    Build a recursively verifiable portable-v3 website snapshot with optional verified-baseline comparison.

  • workflow_runRun an evidence-pack workflow

    Run a registered evidence-pack workflow through workflow.request.v1 and return workflow.result.v1. Supported workflows are website, brand, product, styleguide, visual/image, screenshot, policy, relationship, dataset, fetch, and crawl. Browser use remains explicitly gated.

Compare and alternatives

    Featured in

    Signals

    Stars
    29
    Forks
    1
    Open issues
    0
    Contributors
    7
    Weekly downloads
    348
    Last commit
    Oct 23, 2026 (5 days ago)
    Last release
    Oct 23, 2026 (5 days ago)
    Last package publish
    Oct 23, 2026 (5 days ago)

    From GitHub (fetched Oct 24, 2026) and the package registry.

    Details

    Registry name
    io.github.raintree-technology/docpull
    Version
    6.8.0
    Auth
    No auth
    Packages
    pypi docpull
    License
    MIT
    Repository
    github.com/raintree-technology/docpull
    Website
    github.com/raintree-technology/docpull
    First seen
    Oct 24, 2026
    Last checked
    Oct 24, 2026
    Source
    Official MCP Registrygithub-graphql

    More from raintree-technology

    Cite this page

    DocPull MCP server. Agent Mart. https://agentmart.dev/connectors/raintree-technology-docpull (last updated 2026-10-24)