Pandoc Bridge
Enables document format conversion between 40+ formats including Markdown, HTML, LaTeX, DOCX, and PDF using Pandoc, with support for both text and file-based conversions.
README
Pandoc Bridge
A document format conversion service using Pandoc, supporting both REST (OpenAPI 3.1) and MCP (Model Context Protocol) interfaces.
Features
- 40+ Format Support: Convert between Markdown, HTML, LaTeX, DOCX, PDF, and many more formats
- Dual Protocol: REST API with OpenAPI documentation and MCP server for AI assistants
- Bearer Token Auth: Shared JWT authentication across both protocols
- Streaming Support: Server-Sent Events for progress tracking
- Docker Ready: Production-ready containerization
Quick Start
Using Docker Compose
# Start the service
docker-compose up -d
# Check health
curl http://localhost:8000/health
Local Development
# Install dependencies
pip install -e ".[dev]"
# Run the server
uvicorn src.main:app --reload
# Or use the CLI
pandoc-bridge
API Endpoints
Public Endpoints
| Endpoint | Description |
|---|---|
GET / |
Service information |
GET /health |
Health check with Pandoc version |
GET /api/v1/formats |
List supported formats |
GET /docs |
Swagger UI documentation |
Protected Endpoints (Require Bearer Token)
| Endpoint | Description |
|---|---|
POST /api/v1/convert/text |
Convert text content |
POST /api/v1/convert/file |
Convert uploaded file |
POST /api/v1/convert/stream |
Stream conversion with SSE |
MCP Endpoint
| Endpoint | Description |
|---|---|
POST /mcp |
MCP Streamable HTTP transport |
Usage Examples
Convert Text (REST)
# Get a token (for development, tokens are pre-configured)
TOKEN="your-jwt-token"
# Convert Markdown to HTML
curl -X POST http://localhost:8000/api/v1/convert/text \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{
"content": "# Hello World\n\nThis is **bold** text.",
"from_format": "markdown",
"to_format": "html"
}'
Convert File (REST)
curl -X POST http://localhost:8000/api/v1/convert/file \
-H "Authorization: Bearer $TOKEN" \
-F "file=@document.md" \
-F "to_format=docx" \
-o converted.docx
Using with Claude Desktop (MCP)
Add to your Claude Desktop configuration:
{
"mcpServers": {
"pandoc-bridge": {
"command": "npx",
"args": ["-y", "mcp-remote", "http://localhost:8000/mcp"]
}
}
}
Or with authentication:
{
"mcpServers": {
"pandoc-bridge": {
"command": "npx",
"args": [
"-y", "mcp-remote",
"--header", "Authorization: Bearer ${PANDOC_TOKEN}",
"http://localhost:8000/mcp"
]
}
}
}
MCP Tools
| Tool | Description |
|---|---|
convert_text |
Convert text between formats |
convert_file_base64 |
Convert base64-encoded files |
list_formats |
Get supported formats |
get_service_info |
Get version information |
MCP Resources
| Resource | Description |
|---|---|
formats://list |
JSON list of supported formats |
formats://matrix |
Format conversion compatibility matrix |
Authentication
Scopes
| Scope | Description |
|---|---|
formats:read |
Query formats (public) |
convert:text |
Text conversion |
convert:file |
File conversion |
admin |
Full access |
Token Format
{
"sub": "client-id",
"scopes": ["formats:read", "convert:text", "convert:file"],
"exp": 1735600000,
"iat": 1735513600
}
Configuration
Environment variables:
| Variable | Default | Description |
|---|---|---|
HOST |
0.0.0.0 |
Server host |
PORT |
8000 |
Server port |
LOG_LEVEL |
INFO |
Logging level |
JWT_SECRET_KEY |
(required) | JWT signing key (min 32 chars) |
JWT_EXPIRE_HOURS |
24 |
Token expiration time |
MAX_FILE_SIZE_MB |
50 |
Maximum file size |
CONVERSION_TIMEOUT |
60 |
Conversion timeout in seconds |
Development
Run Tests
# Install dev dependencies
pip install -e ".[dev]"
# Run tests
pytest
# Run with coverage
pytest --cov=src
Lint Code
# Run ruff
ruff check src tests
ruff format src tests
Type Check
mypy src
Architecture
┌─────────────────────────────────────────────────────┐
│ FastAPI App │
│ (Port 8000) │
├─────────────────────┬───────────────────────────────┤
│ REST API │ MCP Server │
│ /api/v1/* │ /mcp │
│ (OpenAPI 3.1) │ (Streamable HTTP) │
├─────────────────────┴───────────────────────────────┤
│ Shared Auth (Bearer Token) │
├─────────────────────────────────────────────────────┤
│ ConversionService (Core) │
├─────────────────────────────────────────────────────┤
│ Pandoc CLI │
└─────────────────────────────────────────────────────┘
License
MIT License - see LICENSE for details.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。