Octagon MCP Server: Real-Time Public Private Market Intelligence

Overview

The Octagon MCP Server is an open-source Model Context Protocol (MCP) server that provides real-time access to public and private market intelligence. It centralizes searchable datasets such as stock metadata, SEC filings (including 13-Fs), private deal records, and investor information so applications and analysts can query, enrich, and surface market signals quickly. The server is intended for developers building research tools, dashboards, alerting systems, and AI agents that need low-latency, contextualized market data.

Designed to be a deployment-ready backend, the MCP server exposes a simple API for full-text and structured queries, integrates multiple data sources, and supports streaming and batch workflows. Because it acts as a model context provider, it can be used alongside language models and analytics pipelines to supply authoritative evidence (filings, ownership records, deal terms) that help avoid hallucinations in automated analysis.

Features

• Unified API for public (stocks, SEC filings) and private (deals, investor details) market data
• Full-text search across filings, news, and deal descriptions
• Structured queries for 13-F positions, ownership changes, and investor profiles
• Fast, scalable endpoints suitable for real-time apps and background workers
• Support for incremental ingestion and re-indexing of new filings and data feeds
• JSON-based responses that are easy to integrate with UIs and ML pipelines
• Health and metrics endpoints for monitoring and observability
• Open-source: source code and issue tracking on GitHub

Installation / Configuration

Clone the repository and follow the basic setup. The project supports running locally for development or in containers for production.

1. Clone the repo

git clone https://github.com/OctagonAI/octagon-mcp-server.git
cd octagon-mcp-server

2. Install dependencies (Node.js example)

# If the project uses npm
npm install

# Or if it uses yarn
yarn install

3. Environment configuration Create a .env file (or use your secrets manager) and set required variables. A minimal example:

PORT=3000
NODE_ENV=production
DATABASE_URL=postgres://user:pass@db:5432/octagon
SEARCH_INDEX_URL=http://localhost:9200
SEC_API_KEY=your_sec_data_provider_key
PRIVATE_DATA_API_KEY=your_private_deals_provider_key

4. Run the server

# Development
npm run dev

# Production
npm start

5. Run with Docker (optional)

docker build -t octagon/mcp-server .
docker run -p 3000:3000 --env-file .env octagon/mcp-server

Note: adjust the exact commands to match the repository’s package scripts. See the repository README for provider-specific ingestion keys and example docker-compose files.

Available Resources

• GitHub repository: https://github.com/OctagonAI/octagon-mcp-server
• API endpoints (examples listed below)
• Ingestion connectors for SEC filings, 13-F feeds, and private deal aggregators
• Health, metrics, and logging endpoints for observability

Example API endpoint summary:

Use Cases

• Research dashboard for equities analysts

  • Use the /filings and /13f endpoints to display recent filings and institutional ownership changes. Combine with search results to surface excerpts and supporting evidence for analyst notes.

• Fund strategy signal generation

  • Run batch queries against the /13f endpoint to identify large position changes, then feed results into a signals pipeline or a backtesting workflow.

• Due diligence for private deals

  • Query /deals and /investors to assemble investor histories, prior round terms, and co-investor networks for fundraising or M&A scouting.

• LLM-powered research assistant

  • Use the MCP server as a context provider when answering questions about a company. Send evidence snippets from filings and 13-Fs alongside prompts to reduce hallucination and improve traceability.

• Alerting and monitoring

  • Subscribe to incremental ingestion events or poll the /filings endpoint to trigger alerts on material disclosures or insider transactions.

Example: Simple Search Request

Query the server for documents mentioning a ticker and a keyword:

curl -X POST "http://localhost:3000/search" \
  -H "Content-Type: application/json" \
  -d '{
    "query": "AAPL buyback",
    "filters": {"type": ["filing","news"], "since": "2025-01-01"},
    "limit": 10
  }'

Response (abridged):

{
  "results": [
    {"id": "filing-1234", "type": "filing", "title": "Form 8-K: Share Repurchase Program", "excerpt": "..."},
    {"id": "news-5678", "type": "news", "title": "Apple expands buyback", "excerpt": "..."}
  ],
  "total": 2
}

Getting Help and Contributing

Refer to the GitHub repository for detailed setup guides, provider-specific ingestion instructions, and contribution guidelines. Open issues and pull requests to report bugs, request features, or propose integrations with new data sources.