Initial commit: Outline MCP server with hybrid light pattern
Some checks failed
Build and Push Docker Image / build (push) Failing after 0s

This commit is contained in:
Ben
2025-12-30 19:36:17 +00:00
parent 6089596202
commit 5da443cef7
9 changed files with 1110 additions and 0 deletions

7
.env.example Normal file
View File

@@ -0,0 +1,7 @@
# Outline MCP Server Configuration
# Outline instance URL (without trailing slash)
OUTLINE_API_URL=https://docs.example.com
# API token from Outline Settings > API Tokens
OUTLINE_API_TOKEN=your_api_token_here

View File

@@ -0,0 +1,35 @@
name: Build and Push Docker Image
on:
push:
branches:
- main
- master
jobs:
build:
runs-on: ubuntu-latest
steps:
- name: Checkout
uses: actions/checkout@v4
- name: Set up Docker Buildx
uses: docker/setup-buildx-action@v3
- name: Login to Container Registry
uses: docker/login-action@v3
with:
registry: gitea.ext.ben.io
username: ${{ secrets.REGISTRY_USERNAME }}
password: ${{ secrets.REGISTRY_PASSWORD }}
- name: Build and push
uses: docker/build-push-action@v5
with:
context: .
push: true
tags: |
gitea.ext.ben.io/b3nw/outline-mcp-custom:latest
gitea.ext.ben.io/b3nw/outline-mcp-custom:${{ github.sha }}
cache-from: type=gha
cache-to: type=gha,mode=max

1
BLUEPRINT.md Symbolic link
View File

@@ -0,0 +1 @@
../BLUEPRINT.md

20
Dockerfile Normal file
View File

@@ -0,0 +1,20 @@
FROM python:3.11-slim
WORKDIR /app
# Install dependencies
COPY pyproject.toml .
RUN pip install --no-cache-dir .
# Copy application
COPY server.py .
# Expose port
EXPOSE 8000
# Health check
HEALTHCHECK --interval=30s --timeout=10s --start-period=5s --retries=3 \
CMD python -c "import httpx; httpx.get('http://localhost:8000/health').raise_for_status()"
# Run server
CMD ["python", "server.py"]

386
IMPLEMENTATION.md Normal file
View File

@@ -0,0 +1,386 @@
# Outline MCP Implementation Guide
This document provides implementation details and comprehensive API reference for AI agents using the Outline MCP server.
## Architecture
The server follows the Hybrid MCP Light pattern:
- 4 specific tools for common read operations
- 1 API passthrough for full coverage
- Embedded API reference resource
## Tools Overview
### search_documents
Full-text search across all documents.
```
Params: query (required), collection_id, limit (default 25), include_archived
Returns: Search results with document titles, IDs, and context snippets
```
### get_document
Retrieve a document by ID.
```
Params: document_id (required), share_id
Returns: Document metadata and full Markdown content
```
### list_collections
List all collections in the workspace.
```
Params: limit (default 25), offset
Returns: Collection names, IDs, permissions, and metadata
```
### list_collection_documents
List documents within a specific collection.
```
Params: collection_id (required), limit, offset
Returns: Document tree structure for the collection
```
### outline_api_call
Raw API passthrough for any Outline endpoint.
```
Params: method (required), params (JSON string)
Returns: Raw API response as JSON string
```
---
## Outline API Reference
Outline uses an RPC-style API where all requests are POST to /api/<method>.
All responses have format: {"ok": true/false, "data": ...}
### Authentication
All requests require Bearer token: Authorization: Bearer <token>
Get your token from Outline Settings > API Tokens.
### Pagination
Most list endpoints accept: limit (default 25, max 100), offset (default 0)
Response includes: pagination.nextPath if more results exist.
---
## Documents
### documents.list
List documents. Params: collectionId?, userId?, parentDocumentId?, sort?, direction?, limit?, offset?
### documents.info
Get document by ID. Params: id (required), shareId?
### documents.search
Full-text search. Params: query (required), collectionId?, userId?, dateFilter?, includeArchived?, includeDrafts?, limit?, offset?
### documents.create
Create document. Params: title (required), collectionId (required), text?, parentDocumentId?, templateId?, template?, publish?
### documents.update
Update document. Params: id (required), title?, text?, append?, publish?, done?
### documents.delete
Delete document. Params: id (required), permanent?
### documents.move
Move document. Params: id (required), collectionId?, parentDocumentId?
### documents.archive
Archive document. Params: id (required)
### documents.unarchive
Restore archived document. Params: id (required)
### documents.restore
Restore deleted document. Params: id (required), revisionId?
### documents.export
Export document. Params: id (required)
Returns Markdown content.
### documents.import
Import document. Multipart form: file, collectionId, parentDocumentId?, template?, publish?
### documents.star
Star a document. Params: id (required)
### documents.unstar
Remove star. Params: id (required)
### documents.templatize
Convert to template. Params: id (required)
### documents.unpublish
Unpublish document. Params: id (required)
---
## Collections
### collections.list
List all collections. Params: limit?, offset?
### collections.info
Get collection details. Params: id (required)
### collections.documents
Get document structure for collection. Params: id (required), limit?, offset?
### collections.create
Create collection. Params: name (required), description?, color?, permission?, sharing?
### collections.update
Update collection. Params: id (required), name?, description?, color?, permission?, sharing?
### collections.delete
Delete collection. Params: id (required)
### collections.add_user
Add user to collection. Params: id (required), userId (required), permission?
### collections.remove_user
Remove user from collection. Params: id (required), userId (required)
### collections.memberships
List collection members. Params: id (required), query?, permission?, limit?, offset?
### collections.add_group
Add group to collection. Params: id (required), groupId (required), permission?
### collections.remove_group
Remove group. Params: id (required), groupId (required)
### collections.group_memberships
List group memberships. Params: id (required), query?, permission?, limit?, offset?
### collections.export
Export entire collection. Params: id (required), format?
### collections.export_all
Export all collections. Params: format?
---
## Users
### users.list
List workspace users. Params: query?, filter?, sort?, direction?, limit?, offset?
Filter options: all, active, invited, suspended, admins
### users.info
Get user by ID. Params: id (required)
### users.invite
Invite users. Params: invites (array of {email, name, role})
### users.update
Update user. Params: id (required), name?, avatarUrl?
### users.delete
Delete user. Params: id (required)
### users.suspend
Suspend user. Params: id (required)
### users.activate
Activate suspended user. Params: id (required)
### users.promote
Promote to admin. Params: id (required)
### users.demote
Remove admin. Params: id (required)
---
## Groups
### groups.list
List groups. Params: query?, sort?, direction?, limit?, offset?
### groups.info
Get group. Params: id (required)
### groups.create
Create group. Params: name (required)
### groups.update
Update group. Params: id (required), name (required)
### groups.delete
Delete group. Params: id (required)
### groups.memberships
List group members. Params: id (required), query?, limit?, offset?
### groups.add_user
Add user to group. Params: id (required), userId (required)
### groups.remove_user
Remove user from group. Params: id (required), userId (required)
---
## Comments
### comments.list
List comments on document. Params: documentId (required), limit?, offset?
### comments.create
Create comment. Params: documentId (required), data (required - ProseMirror JSON)
### comments.update
Update comment. Params: id (required), data (required)
### comments.delete
Delete comment. Params: id (required)
---
## Revisions
### revisions.list
List document revisions. Params: documentId (required), limit?, offset?
### revisions.info
Get revision. Params: id (required)
---
## Shares
### shares.list
List document shares. Params: documentId?, sort?, direction?, limit?, offset?
### shares.info
Get share by ID or documentId. Params: id?, documentId?
### shares.create
Create share link. Params: documentId (required), published?, urlId?, includeChildDocuments?
### shares.update
Update share. Params: id (required), published?, urlId?, includeChildDocuments?
### shares.revoke
Revoke share. Params: id (required)
---
## Stars
### stars.list
List starred documents. Params: limit?, offset?
### stars.create
Star an item. Params: documentId?, collectionId?
### stars.delete
Remove star. Params: id (required)
---
## Events
### events.list
List audit events. Params: name?, actorId?, documentId?, collectionId?, auditLog?, sort?, direction?, limit?, offset?
---
## File Operations
### fileOperations.list
List file operations. Params: type?, limit?, offset?
### fileOperations.info
Get file operation status. Params: id (required)
### fileOperations.redirect
Get download URL. Params: id (required)
### fileOperations.delete
Delete file operation. Params: id (required)
---
## Attachments
### attachments.create
Create attachment upload URL. Params: name (required), documentId (required), size (required), contentType (required)
### attachments.delete
Delete attachment. Params: id (required)
### attachments.redirect
Get attachment URL. Params: id (required)
---
## Auth
### auth.info
Get current user and team info. No params.
### auth.config
Get auth configuration. No params.
---
## Views
### views.list
List document views. Params: documentId (required), limit?, offset?
### views.create
Record a view. Params: documentId (required)
---
## Common Response Patterns
### Success
```json
{"ok": true, "data": {...}, "pagination": {"nextPath": "...", "limit": 25, "offset": 0}}
```
### Error
```json
{"ok": false, "error": "message", "status": 401}
```
### Rate Limiting
HTTP 429 with Retry-After header
---
## Usage Examples
### Create a document
```
method: "documents.create"
params: {"title": "My Document", "collectionId": "collection-uuid", "text": "# Hello\n\nDocument content here.", "publish": true}
```
### Update document content
```
method: "documents.update"
params: {"id": "document-uuid", "text": "# Updated Content\n\nNew content here."}
```
### Delete a document
```
method: "documents.delete"
params: {"id": "document-uuid"}
```
### Invite a user
```
method: "users.invite"
params: {"invites": [{"email": "user@example.com", "name": "User Name", "role": "member"}]}
```
### Create a collection
```
method: "collections.create"
params: {"name": "Engineering", "description": "Engineering documentation", "permission": "read_write"}
```

66
README.md Normal file
View File

@@ -0,0 +1,66 @@
# Outline MCP Server
A lightweight MCP server for interacting with self-hosted [Outline](https://www.getoutline.com/) knowledge bases.
## Features
- **4 specific tools** for common operations (search, get document, list collections)
- **API passthrough** for full Outline API coverage
- **Embedded API reference** resource for agent self-service
- **Docker-ready** with health checks
## Tools
| Tool | Description |
|------|-------------|
| `search_documents` | Full-text search across all documents |
| `get_document` | Retrieve a document by ID with Markdown content |
| `list_collections` | List all collections in the workspace |
| `list_collection_documents` | List documents within a specific collection |
| `outline_api_call` | Raw API passthrough for any Outline endpoint |
## Resources
| URI | Description |
|-----|-------------|
| `outline://api-reference` | Comprehensive Outline API documentation |
## Configuration
Copy `.env.example` to `.env` and configure:
```bash
OUTLINE_API_URL=https://docs.example.com
OUTLINE_API_TOKEN=your_api_token_here
```
Get your API token from Outline: **Settings > API Tokens**
## Running
### Docker Compose
```bash
docker compose up -d
```
### Local Development
```bash
pip install -e .
python server.py
```
## Health Check
```bash
curl http://localhost:8000/health
```
## MCP Endpoint
The MCP server is available at `http://localhost:8000/mcp`
## License
MIT

14
docker-compose.yml Normal file
View File

@@ -0,0 +1,14 @@
services:
outline-mcp:
build: .
ports:
- "${PORT:-8100}:8000"
env_file:
- .env
restart: unless-stopped
healthcheck:
test: ["CMD", "python", "-c", "import httpx; httpx.get('http://localhost:8000/health').raise_for_status()"]
interval: 30s
timeout: 10s
retries: 3
start_period: 5s

22
pyproject.toml Normal file
View File

@@ -0,0 +1,22 @@
[project]
name = "outline-mcp-custom"
version = "0.1.0"
description = "MCP server for Outline knowledge base"
requires-python = ">=3.11"
dependencies = [
"fastmcp>=2.0.0",
"httpx>=0.27.0",
"python-dotenv>=1.0.0",
"starlette>=0.38.0",
"uvicorn>=0.30.0",
]
[project.optional-dependencies]
dev = [
"pytest>=8.0.0",
"pytest-asyncio>=0.23.0",
]
[build-system]
requires = ["hatchling"]
build-backend = "hatchling.build"

559
server.py Normal file
View File

@@ -0,0 +1,559 @@
import os
import json
import logging
import httpx
from fastmcp import FastMCP
from starlette.applications import Starlette
from starlette.responses import JSONResponse
from starlette.routing import Route, Mount
from dotenv import load_dotenv
# Load environment variables
load_dotenv()
# Logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# --- Configuration ---
OUTLINE_API_URL = os.getenv("OUTLINE_API_URL", "").rstrip("/")
OUTLINE_API_TOKEN = os.getenv("OUTLINE_API_TOKEN", "")
class OutlineClient:
"""HTTP client for Outline API with authentication."""
def __init__(self, base_url: str, api_token: str):
self.base_url = base_url.rstrip("/")
self.api_token = api_token
self._client: httpx.AsyncClient | None = None
@property
def client(self) -> httpx.AsyncClient:
if self._client is None or self._client.is_closed:
self._client = httpx.AsyncClient(
base_url=self.base_url,
headers={
"Authorization": f"Bearer {self.api_token}",
"Content-Type": "application/json",
},
timeout=30.0,
)
return self._client
async def call(self, method: str, params: dict | None = None) -> dict:
"""Execute an Outline API call.
Args:
method: API method (e.g., 'documents.list', 'collections.info')
params: Optional parameters to send in request body
Returns:
API response as dict
"""
if params is None:
params = {}
try:
response = await self.client.post(f"/api/{method}", json=params)
response.raise_for_status()
return response.json()
except httpx.HTTPStatusError as e:
error_body = e.response.text
try:
error_json = e.response.json()
return {
"ok": False,
"error": error_json.get("message", str(e)),
"status": e.response.status_code,
}
except Exception:
return {
"ok": False,
"error": error_body or str(e),
"status": e.response.status_code,
}
except httpx.RequestError as e:
return {"ok": False, "error": f"Request failed: {str(e)}"}
async def check_health(self) -> tuple[bool, str]:
"""Check API connectivity by calling auth.info."""
try:
result = await self.call("auth.info")
if result.get("ok"):
return True, "Connected"
return False, result.get("error", "Unknown error")
except Exception as e:
return False, str(e)
async def close(self):
if self._client and not self._client.is_closed:
await self._client.aclose()
# Initialize client
outline_client = OutlineClient(OUTLINE_API_URL, OUTLINE_API_TOKEN)
# --- FastMCP Server ---
mcp = FastMCP("Outline MCP")
@mcp.tool()
async def search_documents(
query: str,
collection_id: str = None,
limit: int = 25,
include_archived: bool = False,
) -> str:
"""Search for documents in Outline using full-text search.
Args:
query: Search query string
collection_id: Optional collection ID to search within
limit: Maximum number of results (default 25, max 100)
include_archived: Include archived documents in results
Returns:
JSON string with search results containing document titles, IDs, and context
"""
params = {
"query": query,
"limit": min(limit, 100),
"includeArchived": include_archived,
}
if collection_id:
params["collectionId"] = collection_id
result = await outline_client.call("documents.search", params)
return json.dumps(result, indent=2)
@mcp.tool()
async def get_document(document_id: str, share_id: str = None) -> str:
"""Retrieve a document by ID, returning its full Markdown content.
Args:
document_id: The document UUID
share_id: Optional share ID for accessing shared documents
Returns:
JSON string with document metadata and Markdown content
"""
params = {"id": document_id}
if share_id:
params["shareId"] = share_id
result = await outline_client.call("documents.info", params)
return json.dumps(result, indent=2)
@mcp.tool()
async def list_collections(limit: int = 25, offset: int = 0) -> str:
"""List all collections (folders) in the Outline workspace.
Args:
limit: Maximum number of results (default 25)
offset: Pagination offset
Returns:
JSON string with collection names, IDs, and metadata
"""
params = {"limit": limit, "offset": offset}
result = await outline_client.call("collections.list", params)
return json.dumps(result, indent=2)
@mcp.tool()
async def list_collection_documents(
collection_id: str, limit: int = 25, offset: int = 0
) -> str:
"""List all documents within a specific collection.
Args:
collection_id: The collection UUID
limit: Maximum number of results (default 25)
offset: Pagination offset
Returns:
JSON string with document tree structure for the collection
"""
params = {"id": collection_id, "limit": limit, "offset": offset}
result = await outline_client.call("collections.documents", params)
return json.dumps(result, indent=2)
@mcp.tool()
async def outline_api_call(method: str, params: str = "{}") -> str:
"""Execute a raw Outline API call for any endpoint.
Use the 'outline://api-reference' resource to discover available methods
and their parameters.
Args:
method: API method name (e.g., 'documents.create', 'users.list')
params: JSON string of parameters to send in request body
Returns:
JSON string with raw API response
Examples:
- method='documents.create', params='{"title": "New Doc", "collectionId": "uuid"}'
- method='users.list', params='{"limit": 10}'
- method='documents.delete', params='{"id": "document-uuid"}'
"""
try:
parsed_params = json.loads(params) if params else {}
except json.JSONDecodeError as e:
return json.dumps({"ok": False, "error": f"Invalid JSON params: {str(e)}"})
result = await outline_client.call(method, parsed_params)
return json.dumps(result, indent=2)
# --- API Reference Resource ---
API_REFERENCE = """# Outline API Reference
Outline uses an RPC-style API where all requests are POST to /api/<method>.
All responses have format: {"ok": true/false, "data": ...}
## Authentication
All requests require Bearer token: Authorization: Bearer <token>
Get your token from Outline Settings > API Tokens.
## Pagination
Most list endpoints accept: limit (default 25, max 100), offset (default 0)
Response includes: pagination.nextPath if more results exist.
---
## Documents
### documents.list
List documents. Params: collectionId?, userId?, parentDocumentId?, sort?, direction?, limit?, offset?
### documents.info
Get document by ID. Params: id (required), shareId?
### documents.search
Full-text search. Params: query (required), collectionId?, userId?, dateFilter?, includeArchived?, includeDrafts?, limit?, offset?
### documents.create
Create document. Params: title (required), collectionId (required), text?, parentDocumentId?, templateId?, template?, publish?
### documents.update
Update document. Params: id (required), title?, text?, append?, publish?, done?
### documents.delete
Delete document. Params: id (required), permanent?
### documents.move
Move document. Params: id (required), collectionId?, parentDocumentId?
### documents.archive
Archive document. Params: id (required)
### documents.unarchive
Restore archived document. Params: id (required)
### documents.restore
Restore deleted document. Params: id (required), revisionId?
### documents.export
Export document. Params: id (required)
Returns Markdown content.
### documents.import
Import document. Multipart form: file, collectionId, parentDocumentId?, template?, publish?
### documents.star
Star a document. Params: id (required)
### documents.unstar
Remove star. Params: id (required)
### documents.templatize
Convert to template. Params: id (required)
### documents.unpublish
Unpublish document. Params: id (required)
---
## Collections
### collections.list
List all collections. Params: limit?, offset?
### collections.info
Get collection details. Params: id (required)
### collections.documents
Get document structure for collection. Params: id (required), limit?, offset?
### collections.create
Create collection. Params: name (required), description?, color?, permission?, sharing?
### collections.update
Update collection. Params: id (required), name?, description?, color?, permission?, sharing?
### collections.delete
Delete collection. Params: id (required)
### collections.add_user
Add user to collection. Params: id (required), userId (required), permission?
### collections.remove_user
Remove user from collection. Params: id (required), userId (required)
### collections.memberships
List collection members. Params: id (required), query?, permission?, limit?, offset?
### collections.add_group
Add group to collection. Params: id (required), groupId (required), permission?
### collections.remove_group
Remove group. Params: id (required), groupId (required)
### collections.group_memberships
List group memberships. Params: id (required), query?, permission?, limit?, offset?
### collections.export
Export entire collection. Params: id (required), format?
### collections.export_all
Export all collections. Params: format?
---
## Users
### users.list
List workspace users. Params: query?, filter?, sort?, direction?, limit?, offset?
Filter options: all, active, invited, suspended, admins
### users.info
Get user by ID. Params: id (required)
### users.invite
Invite users. Params: invites (array of {email, name, role})
### users.update
Update user. Params: id (required), name?, avatarUrl?
### users.delete
Delete user. Params: id (required)
### users.suspend
Suspend user. Params: id (required)
### users.activate
Activate suspended user. Params: id (required)
### users.promote
Promote to admin. Params: id (required)
### users.demote
Remove admin. Params: id (required)
---
## Groups
### groups.list
List groups. Params: query?, sort?, direction?, limit?, offset?
### groups.info
Get group. Params: id (required)
### groups.create
Create group. Params: name (required)
### groups.update
Update group. Params: id (required), name (required)
### groups.delete
Delete group. Params: id (required)
### groups.memberships
List group members. Params: id (required), query?, limit?, offset?
### groups.add_user
Add user to group. Params: id (required), userId (required)
### groups.remove_user
Remove user from group. Params: id (required), userId (required)
---
## Comments
### comments.list
List comments on document. Params: documentId (required), limit?, offset?
### comments.create
Create comment. Params: documentId (required), data (required - ProseMirror JSON)
### comments.update
Update comment. Params: id (required), data (required)
### comments.delete
Delete comment. Params: id (required)
---
## Revisions
### revisions.list
List document revisions. Params: documentId (required), limit?, offset?
### revisions.info
Get revision. Params: id (required)
---
## Shares
### shares.list
List document shares. Params: documentId?, sort?, direction?, limit?, offset?
### shares.info
Get share by ID or documentId. Params: id?, documentId?
### shares.create
Create share link. Params: documentId (required), published?, urlId?, includeChildDocuments?
### shares.update
Update share. Params: id (required), published?, urlId?, includeChildDocuments?
### shares.revoke
Revoke share. Params: id (required)
---
## Stars
### stars.list
List starred documents. Params: limit?, offset?
### stars.create
Star an item. Params: documentId?, collectionId?
### stars.delete
Remove star. Params: id (required)
---
## Events
### events.list
List audit events. Params: name?, actorId?, documentId?, collectionId?, auditLog?, sort?, direction?, limit?, offset?
---
## File Operations
### fileOperations.list
List file operations. Params: type?, limit?, offset?
### fileOperations.info
Get file operation status. Params: id (required)
### fileOperations.redirect
Get download URL. Params: id (required)
### fileOperations.delete
Delete file operation. Params: id (required)
---
## Attachments
### attachments.create
Create attachment upload URL. Params: name (required), documentId (required), size (required), contentType (required)
### attachments.delete
Delete attachment. Params: id (required)
### attachments.redirect
Get attachment URL. Params: id (required)
---
## Auth
### auth.info
Get current user and team info. No params.
### auth.config
Get auth configuration. No params.
---
## Views
### views.list
List document views. Params: documentId (required), limit?, offset?
### views.create
Record a view. Params: documentId (required)
---
## Common Response Patterns
Success:
{"ok": true, "data": {...}, "pagination": {"nextPath": "...", "limit": 25, "offset": 0}}
Error:
{"ok": false, "error": "message", "status": 401}
Rate Limiting:
HTTP 429 with Retry-After header
"""
@mcp.resource("outline://api-reference")
def get_api_reference() -> str:
"""Returns comprehensive Outline API documentation for use with outline_api_call tool."""
return API_REFERENCE
# --- Health Check Endpoint ---
async def health(request):
"""Health check endpoint for Docker/load balancer."""
if not OUTLINE_API_URL or not OUTLINE_API_TOKEN:
return JSONResponse(
{
"status": "degraded",
"error": "Missing OUTLINE_API_URL or OUTLINE_API_TOKEN",
},
status_code=503,
)
healthy, message = await outline_client.check_health()
if healthy:
return JSONResponse({"status": "ok"})
return JSONResponse({"status": "degraded", "error": message}, status_code=503)
# --- ASGI Application ---
def create_app():
"""Create the ASGI application with health check and MCP routes."""
mcp_app = mcp.http_app()
routes = [
Route("/health", health, methods=["GET"]),
Mount("/", app=mcp_app),
]
return Starlette(routes=routes, lifespan=mcp_app.lifespan)
# Create the app instance for uvicorn
app = create_app()
if __name__ == "__main__":
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=8000)