@cyanheads/docgen-mcp-server
Render HTML/markdown to PDF, export rows to xlsx, and fill AcroForm PDFs via MCP. STDIO or Streamable HTTP.
README
<div align="center"> <h1>@cyanheads/docgen-mcp-server</h1> <p><b>Render HTML/markdown to PDF, export rows to xlsx, and fill AcroForm PDFs via MCP. STDIO or Streamable HTTP.</b> <div>4 Tools • 1 Resource</div> </p> </div>
<div align="center">
</div>
<div align="center">
</div>
Why docgen
An agent can write a perfect invoice's HTML, a clean data table, or a filled-form field map as tokens — but it cannot emit bytes. docgen is the renderer that closes the gap: structured content in, a downloadable binary document out. It wraps no external API; the "service" is a bundled rendering stack (pdf-lib for PDF and form fill, exceljs for spreadsheets, marked for the markdown path). Every output is stored tenant-scoped with a short TTL and handed back as a stable resource URI plus inline base64 when small enough.
Tools
Four tools sharing one delivery shape — three renderers that write a stored document and one read that re-fetches by id. Every render/export/fill call returns a DocumentEnvelope (a documentId, a docgen://document/{id} resource URI, byte size, TTL, and inline base64 when the artifact is small enough); docgen_get_document returns the same envelope for an id you already hold.
| Tool | Description |
|---|---|
docgen_render_pdf |
Render HTML, markdown, or a {{key}} template + data object to a downloadable PDF. |
docgen_export_spreadsheet |
Render one or more named worksheets of row objects to a downloadable .xlsx workbook. |
docgen_fill_form |
Fill the AcroForm fields of a supplied PDF (base64 or https URL) and optionally flatten it. |
docgen_get_document |
Re-fetch a previously rendered document by the id a render/export/fill tool returned. |
docgen_render_pdf
Render content to a downloadable PDF.
- Provide exactly one
source:{ html }(raw HTML you compose — the recommended path),{ markdown }(converted to HTML, then rendered), or{ template, data }(a{{key}}template filled from a data object, with server-owned layout) pageOptionscontrol size (A4/Letter/Legal/A3/A5, defaultLetter), orientation, per-side margins (CSS lengths like"10mm","0.5in","72pt"), header/footer text (supporting the{{page}},{{total}},{{date}}tokens), and automatic page numbers- The lightweight engine renders structured layout (headings, paragraphs, lists, tables) but not arbitrary CSS — a
degradedenrichment flag is set when unsupported styling is dropped, so the agent isn't misled about fidelity - Returns a
DocumentEnvelopewithpageCount
docgen_export_spreadsheet
Render tabular data to an .xlsx workbook — the natural export stage for rows pulled from another server.
- Each entry in
sheets[]is a worksheetnameplus an array of row objects (property → scalar string / number / boolean / null) - An optional per-column spec sets the header label, value
type(string/number/date/boolean), column width, and an Excel number/date format string (e.g."#,##0.00","yyyy-mm-dd"); when omitted, columns derive from the first row's keys - An empty
rowsarray yields a header-only sheet; an emptysheets[]is rejected (empty_workbook) - Returns a
DocumentEnvelopewithsheetCount
docgen_fill_form
Fill the AcroForm fields of a supplied PDF and optionally flatten it.
- Provide the source as exactly one of
{ base64 }or{ url }— an https URL fetched behind an SSRF guard (private/loopback/link-local ranges blocked,application/pdfrequired, response size capped); base64 avoids the fetch entirely fieldsis an AcroForm field name → value map; names must match the PDF's internal field names exactly (case-sensitive), obtained from whoever supplied the form (docgen does not expose them)- Names with no AcroForm counterpart come back in
unmatchedFields[]rather than failing the call — correct them and re-render - Set
flatten: trueto bake the values in so the result is no longer editable - AcroForm only — XFA-based PDFs (some government forms) report
not_a_form - Returns a
DocumentEnvelopewithpageCount, plusunmatchedFields[]
docgen_get_document
Re-fetch a stored document by id.
- The
documentIdis obtainable only from an earlierdocgen_render_pdf,docgen_export_spreadsheet, ordocgen_fill_formresult — it is not guessable or constructible - Use it to recover a document whose inline copy was dropped (over the inline size limit) while it is still within its TTL
- An expired or unknown id returns
document_expired; ids are single-render and not reusable
Resources
| Type | Name | Description |
|---|---|---|
| Resource | docgen://document/{documentId} |
A rendered document by id — the raw bytes as a blob (with the document's real mime type) plus a JSON metadata block. Readable until the document TTL expires. |
The resource is the stable-URI delivery surface for hosts that support resources. All document data is also reachable via the tool surface — docgen_get_document is the tool-only twin of this resource, reading the same store and returning the same envelope. Neither re-renders.
Features
Built on @cyanheads/mcp-ts-core:
- Declarative tool and resource definitions — single file per primitive, framework handles registration and validation
- Unified error handling — handlers throw, framework catches, classifies, and formats
- Pluggable auth:
none,jwt,oauth - Swappable storage backends:
in-memory,filesystem,Supabase,Cloudflare KV/R2/D1 - Structured logging with optional OpenTelemetry tracing
- STDIO and Streamable HTTP transports
docgen-specific:
- No external API — a bundled rendering stack (
pdf-lib,exceljs,marked), so renders are local and deterministic with no upstream to fail - One shared
DocumentEnvelopeacross all four tools — the three writers and the reader are interchangeable to the agent, and the resource-URI vs. inline-base64 delivery is decided in one place - Bounded renders — a per-document byte ceiling (
DOCGEN_MAX_DOCUMENT_BYTES) and a wall-clock timeout (DOCGEN_RENDER_TIMEOUT_MS) turn a runaway render into a typed, recoverable error instead of a hang - Tenant-scoped, TTL-bounded storage — a document id minted for one tenant resolves only for that tenant; outputs are downloads, not records, so they expire rather than accumulate
- SSRF-guarded form fetch —
docgen_fill_formwith a URL source resolves DNS and checks the destination IP before fetching, blocking private/loopback/link-local ranges
Agent-friendly output:
- Dual-surface delivery — every envelope field lands in both
structuredContentand theformat()markdown twin, so tool-only and resource-only clients both see thedocumentId, resource URI, inline-availability status, size, and TTL - Inline-vs-resource by size —
inlineBase64is populated only at or underDOCGEN_INLINE_MAX_BYTES, so a large workbook isn't base64-inlined into a tool result; above the threshold, delivery is via the resource URI - Partial-fill reporting —
docgen_fill_formreturnsunmatchedFields[]so the agent learns which field names didn't land and can correct and re-render rather than assuming a clean fill - Typed error contract with recovery hints — each tool declares its failure surface (
invalid_source,template_render_failed,document_too_large,render_timeout,not_a_form,source_unfetchable,document_expired, …) with actionable next-step text
Getting started
Add the following to your MCP client configuration file. No API keys are required.
{
"mcpServers": {
"docgen-mcp-server": {
"type": "stdio",
"command": "bunx",
"args": ["@cyanheads/docgen-mcp-server@latest"],
"env": {
"MCP_TRANSPORT_TYPE": "stdio",
"MCP_LOG_LEVEL": "info"
}
}
}
}
Or with npx (no Bun required):
{
"mcpServers": {
"docgen-mcp-server": {
"type": "stdio",
"command": "npx",
"args": ["-y", "@cyanheads/docgen-mcp-server@latest"],
"env": {
"MCP_TRANSPORT_TYPE": "stdio",
"MCP_LOG_LEVEL": "info"
}
}
}
}
Or with Docker:
{
"mcpServers": {
"docgen-mcp-server": {
"type": "stdio",
"command": "docker",
"args": [
"run", "-i", "--rm",
"-e", "MCP_TRANSPORT_TYPE=stdio",
"ghcr.io/cyanheads/docgen-mcp-server:latest"
]
}
}
}
For Streamable HTTP, set the transport and start the server:
MCP_TRANSPORT_TYPE=http MCP_HTTP_PORT=3010 bun run start:http
# Server listens at http://localhost:3010/mcp
Documents are delivered by the docgen://document/{id} resource (and inline base64 when small enough) over every transport. The envelope's downloadUrl field is reserved for a future HTTP download route and is not emitted in this version.
Prerequisites
- Bun v1.3.2 or higher (or Node.js v24+).
Installation
- Clone the repository:
git clone https://github.com/cyanheads/docgen-mcp-server.git
- Navigate into the directory:
cd docgen-mcp-server
- Install dependencies:
bun install
- Configure environment:
cp .env.example .env
# edit .env to override any defaults
Configuration
All configuration is optional — docgen runs with no required environment variables.
| Variable | Description | Default |
|---|---|---|
DOCGEN_DOCUMENT_TTL_SECONDS |
How long a rendered document is retrievable before it expires, in seconds. | 900 |
DOCGEN_MAX_DOCUMENT_BYTES |
Hard ceiling on a single rendered artifact in bytes; exceeding it aborts the render. | 26214400 |
DOCGEN_RENDER_TIMEOUT_MS |
Per-render wall-clock budget in milliseconds; exceeding it aborts the render. | 30000 |
DOCGEN_INLINE_MAX_BYTES |
Artifacts at or under this byte size are returned inline as base64; larger ones omit it. | 5242880 |
DOCGEN_PDF_ENGINE |
PDF rendering engine. Only lightweight is implemented; chromium is reserved and rejected at startup. |
lightweight |
MCP_TRANSPORT_TYPE |
Transport: stdio or http. |
stdio |
MCP_HTTP_PORT |
Port for the HTTP server. | 3010 |
MCP_AUTH_MODE |
Auth mode: none, jwt, or oauth. |
none |
MCP_PUBLIC_URL |
Public origin behind a TLS proxy. (The downloadUrl envelope field is reserved for a future HTTP download route and is not emitted in this version.) |
— |
MCP_LOG_LEVEL |
Log level (RFC 5424). | info |
STORAGE_PROVIDER_TYPE |
Storage backend for document bytes + metadata. | in-memory |
OTEL_ENABLED |
Enable OpenTelemetry instrumentation (spans, metrics, completion logs). | false |
See .env.example for the full list of optional overrides.
Documents are stored in
ctx.state, which the in-memory provider keeps in process memory — a restart drops every stored document, and an id minted before the restart returnsdocument_expired. This is intended (outputs are downloads, not records); for durable retention across restarts, pointSTORAGE_PROVIDER_TYPEat a persistent backend.
Running the server
Local development
-
Build and run:
# One-time build bun run rebuild # Run the built server bun run start:stdio # or bun run start:http -
Run checks and tests:
bun run devcheck # Lint, format, typecheck, security, changelog sync bun run test # Vitest test suite bun run lint:mcp # Validate MCP definitions against spec
Docker
docker build -t docgen-mcp-server .
docker run --rm -e MCP_TRANSPORT_TYPE=http -p 3010:3010 docgen-mcp-server
The Dockerfile defaults to HTTP transport, stateless session mode, and logs to /var/log/docgen-mcp-server. OpenTelemetry peer dependencies are installed by default — build with --build-arg OTEL_ENABLED=false to omit them.
Project structure
| Directory | Purpose |
|---|---|
src/index.ts |
createApp() entry point — registers the tools/resource and inits the render + storage services. |
src/config |
Server-specific environment variable parsing and validation with Zod. |
src/mcp-server/tools |
Tool definitions (*.tool.ts). |
src/mcp-server/resources |
Resource definitions (*.resource.ts). |
src/services/document |
The rendering stack (RenderService) and artifact store (DocumentStore), shared types, and the SSRF fetch guard. |
tests/ |
Unit and integration tests mirroring src/. |
Development guide
See CLAUDE.md/AGENTS.md for development guidelines and architectural rules. The short version:
- Handlers throw, framework catches — no
try/catchin tool logic - Use
ctx.logfor request-scoped logging,ctx.statefor tenant-scoped storage - Register new tools and resources in
src/index.ts'screateApp()arrays - One shared
DocumentEnvelopeacross all delivery tools — keep the writers and reader interchangeable
Contributing
Issues and pull requests are welcome. Run checks and tests before submitting:
bun run devcheck
bun run test
License
Apache-2.0 — see LICENSE for details.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.