ZBS Index What actually exists in applied AI, with the source next to it

Document understanding

Everything here was classified as document understanding by keyword match against the maintainer's own description, so treat the grouping as a starting point rather than a verdict.

The list is ordered by the most recent commit, not by stars. A popular project that stopped in 2024 is not a better answer than a smaller one shipped last week.

Worked on in the last 90 days (60)

Someone pushed a commit recently. This is the shortlist worth trying first.

  • GEML — a plain-text document format built to be edited in place, one section at a time — last commit 2026-08-08, v1.7.1
    Patch a single block of a GEML document by id. Bad writes are refused; any one block rolls back.
  • esa — last commit 2026-08-08, v0.13.0
    Official MCP server for esa.io - a documentation sharing service for teams
  • io.github.decantr-ai/mcp-server — last commit 2026-08-07, v3.9.4
    AI frontend governance for Contract context, graph artifacts, drift evidence, and repair prompts.
  • io.github.devemberx/mcp-server-polarion — last commit 2026-08-07, v2.0.0
    MCP server for Polarion ALM — read and write documents and work items
  • Aspicio — last commit 2026-08-07, v0.12.0
    DXF and PDF/X-4 for AI agents: structured facts, PNG renders, an interactive in-chat viewer.
  • ellmos FileCommander — last commit 2026-08-07, v1.9.3
    Filesystem MCP server: files, processes, sessions, async search, OCR, archives, safe delete.
  • io.github.DingDawg-dev/dingdawg-healthcare-agent — last commit 2026-08-06, v2.0.4
    HIPAA-aware healthcare AI. Clinical documentation, compliance checks, audit prep. Free local tier.
  • io.github.dingdawg/dingdawg-healthcare-agent — last commit 2026-08-06, v2.0.7
    HIPAA-aware healthcare AI. Clinical documentation, compliance checks, audit prep. Free local tier.
  • NFT Metadata API — last commit 2026-08-06, v1.1.0
    Fetch NFT metadata — name, image, traits from contract+tokenId. x402 micropayment.
  • Doc Bridge — last commit 2026-08-06, v1.2.6
    Resolve documentation ownership, edit roots, checks, and human handoffs before an agent edits code.
  • com.redpanda/docs-mcp — last commit 2026-08-06, v2026.803.133
    Get authoritative answers about Redpanda documentation and search API references.
  • ArcadeDB MCP Server — last commit 2026-08-06, v26.4.1-SNAPSHOT
    Built-in MCP server for ArcadeDB multi-model database (graph, document, vector, time-series)
  • PandaDoc MCP Server — last commit 2026-08-06, v1.0.0
    PandaDoc MCP server for creating, sending, signing, and tracking PandaDoc documents.
  • ArtifactGuard Signed AI Deliverable QA — last commit 2026-08-06, v1.0.0
    Signed PASS/WARN/FAIL QA receipts for AI-generated JSON, ZIP, PDF, and DOCX deliverables.
  • io.github.bekirdag/docdex — last commit 2026-08-06, v0.2.95
    Fast per-repo documentation indexer and full-text search MCP server for codebases.
  • GitHits — last commit 2026-08-06, v0.7.0
    Search public open-source code, documentation, metadata, vulnerabilities, changelogs, and examples.
  • oxidize-pdf MCP Server — last commit 2026-08-06, v0.16.0
    MCP server for PDF generation, parsing, splitting, merging, and manipulation via oxidize-pdf
  • FlexOrch — last commit 2026-08-06, v0.1.2
    Classify documents, extract structured fields, mask PII, export JSONL/RAG datasets for AI agents.
  • ChangeThisFile — last commit 2026-08-06, v1.1.0
    Convert files between 690+ formats: image, video, audio, documents, ebooks, archives. Free, no auth.
  • Bankstatemently — last commit 2026-08-06, v0.4.0
    Parse PDF bank statements into structured transactions, accounts, and balances.
  • Brandsystem MCP — Brand Guidelines for AI — last commit 2026-08-06, v0.16.1
    Use existing brand guidelines with AI. Turn websites, PDFs, and Figma into a .brand runtime.
  • io.github.Custodia-Admin/pagebolt — last commit 2026-08-06, v1.8.2
    Screenshots, PDFs, OG images, page inspection, and narrated video recording for Claude and Cursor.
  • Carbone MCP — last commit 2026-08-06, v1.5.0
    Generate PDF/DOCX/XLSX/PPTX from templates+JSON. Convert Office/HTML/MD to PDF. Universal templating
  • DataCharter — last commit 2026-08-06, v0.23.7
    Local-first, contract-governed data explorer with read-only, PII-masked query tools for AI agents.
  • dev.avakit/avalanche — last commit 2026-08-06, v0.1.25
    Act on Avalanche from your agent: scaffold a dapp, deploy a contract, mint, and read chain state.
  • io.github.chrischall/crowntowncompost-mcp — last commit 2026-08-06, v0.3.0
    Customer access to the Crown Town Compost portal: pickups, invoices, and skips (AI-built).
  • io.github.chrischall/honeybook-mcp — last commit 2026-08-06, v0.4.4
    HoneyBook client-portal MCP server for Claude — view contracts and invoices from wedding vendors
  • Laravel MCP Companion — last commit 2026-08-06, v0.13.0
    Laravel documentation aggregator: multi-version core docs, services, packages, learning paths.
  • com.esti-maite/construction-takeoff-estimating — last commit 2026-08-05, v1.0.0
    Construction takeoff and estimating for AI agents. Measure a drawing PDF, export a priced estimate.
  • io.github.dingdawg/dingdawg-legal-agent — last commit 2026-08-05, v2.0.7
    AI contract review, clause drafting, compliance checklists & legal research.
  • KIAgent — last commit 2026-08-05, v0.53.0
    Your mail, chats and documents indexed into a local SQLite corpus, served to AI clients over MCP.
  • io.github.br9704/aethereum — last commit 2026-08-05, v0.6.0
    The team layer for AI coding agents: shared contracts, collision alerts, E2EE sessions.
  • Mac-letterhead — last commit 2026-08-05, v0.24.2
    Letterhead PDF as digital stationery: prints Markdown without overprinting header, footer, or logo.
  • Docmancer — last commit 2026-08-05, v0.9.14
    Local-only MCP server for source-attributed Markdown memory and documentation retrieval.
  • Fonto Docs — last commit 2026-08-05, v1.6.1
    Fonto (FontoXML) documentation for AI tools. Converts DITA XML to Markdown on demand.
  • io.github.enzoemir1/invoiceflow-mcp — last commit 2026-08-05, v1.4.0
    AI-powered invoice automation. Create invoices, track payments, and manage cash flow.
  • Soracom Knowledge — last commit 2026-08-05, v1.1.4
    Search SORACOM documentation: service guides, FAQ, API references, IoT recipes, etc.
  • KeyHalve - verify sealed documents — last commit 2026-08-05, v1.0.1
    Free, no-account verification of KeyHalve-sealed documents. Read-only; never receives keys.
  • Jobkeepr — last commit 2026-08-04, v1.0.0
    Manage jobs, customers, scheduling, estimates and invoices for a field service business.
  • Vector (Apex Fusion) — last commit 2026-08-04, v0.1.0
    24 keyless tools for AI agents on Vector, the Apex Fusion eUTXO L2: wallets, contracts, registry.
  • zzop — last commit 2026-08-04, v0.29.1
    Deterministic cross-repo contract analysis for AI agents: frontend calls vs backend endpoints.
  • Haypile — last commit 2026-08-04, v0.3.0
    Hybrid search over your local documents, with file and page citations. One binary, fully local.
  • io.github.cloud1map/agent-relay — last commit 2026-08-04, v0.1.3
    Agent Relay MCP: invite-based channels, agent contracts and per-side human approvals.
  • io.github.aicayzer/companies-house — last commit 2026-08-04, v4.0.0
    UK Companies House public register: search companies, officers, ownership, filings and documents.
  • RentRollAPI — last commit 2026-08-04, v1.2.0
    Extract verified data from CRE rent rolls and T12 operating statements (PDF, Excel, CSV, scans)
  • com.pdfgate/mcp-server — last commit 2026-08-04, v1.0.3
    MCP server for the PDFGate API. Generate PDFs, manage documents and handle e-signatures.
  • com.vaiz/mcp — last commit 2026-08-04, v1.1.2
    Connect AI assistants to your Vaiz workspace: tasks, projects, boards, documents and milestones.
  • acikpoz — last commit 2026-08-04, v0.2.0
    Turkish public construction unit-price (birim fiyat) catalog PDFs into structured pozes
  • Caelune — last commit 2026-08-04, v0.5.5
    Read-only local-first MCP server for private Markdown, PDF, and Tika-backed search on Windows.
  • OmniClip RAG — last commit 2026-08-04, v0.4.13
    Read-only local-first MCP server for private Markdown, PDF, and Tika-backed search on Windows.
  • Dashboardbase — last commit 2026-08-04, v1.0.0
    Validate Dashboardbase widget endpoint responses and setup files against the live contract.
  • ca.bluecraneworks/forms — last commit 2026-08-04, v1.1.0
    Canadian contractor forms + tax tools: 126 cited forms, sourced 2026 tax estimates by province.
  • Evidrift — last commit 2026-08-04, v0.4.0
    Catch TypeScript API and OpenAPI contract drift in AI-generated code, then revalidate it in CI.
  • Rendex: Rendering API for Images, PDFs & Content Extraction — last commit 2026-08-04, v1.8.2
    Render HTML, Markdown, or URLs to images, PDF, or branded artifacts; extract and watch pages.
  • Mini Accountant — last commit 2026-08-03, v1.0.0
    Hosted MCP server for Mini Accountant: invoices, expenses, customers, analytics, tax estimates.
  • Screenshot Scout — last commit 2026-08-03, v0.1.1
    Capture screenshots of webpages as images or PDFs with Screenshot Scout.
  • ch.martinelli/jooq-mcp — last commit 2026-08-03, v1.0.0
    An MCP server that provides access to the jOOQ documentation
  • ai.adeu/adeu — last commit 2026-08-03, v1.7.1
    Automated DOCX Redlining Engine
  • io.github.cliwant/mcp-sam-gov — last commit 2026-08-03, v1.12.0
    Keyless MCP: 150 tools for US federal + state/local (SLED) contracting, spending & regulation
  • io.github.DrBaher/compare-cli-mcp — last commit 2026-08-03, v0.1.4
    MCP server for compare-cli: contract version diff/redline and a drift gate as agent tools.

Page 1 of 17

How this page is ordered

Entries are grouped by whether anyone is still working on them, using the date of the most recent push to the repository. They are not ordered by stars, because a star is a bookmark somebody left once and never took back.

Where we have not checked an entry yet, it says so rather than being mixed in with the verified ones.