MOEX ISS and risk analytics MCP
MOEX Market Analyst Agent (OpenRouter + MCP)
Two MCP servers for MOEX data and portfolio risk from a multi-agent MOEX analyst project: quotes, VaR and CFO reports
Low risk
We rate an entry low when it mostly gives the agent instructions and reference material.
Why this level
- Only reads market data and computes analytics, never places orders
- Does not connect to brokerage accounts
Install
Manual install
docker build -t moex-iss-mcp:local -f moex_iss_mcp/Dockerfile . && docker run -p 8000:8000 --env-file env.example moex-iss-mcp:localBuild and run the MOEX data server separately from the rest of the stack.
This is third-party code. Review the repository files before installing.
What it does
The repository ships a demo multi-agent MOEX analyst on OpenRouter, but inside it are two standalone MCP servers with their own Dockerfile and README that can run separately from the rest of the stack. moex_iss_mcp proxies the MOEX ISS API and offers get_security_snapshot, get_ohlcv_timeseries and get_index_constituents_metrics with a configurable requests-per-second limit. risk_analytics_mcp computes portfolio risk and stress scenarios: basic risk metrics and VaR, a ticker correlation matrix, peer search and comparison by index or sector, a deterministic rebalancing suggestion, a CFO liquidity report and index tail metrics. Both servers run over streamable-http with /mcp and /health endpoints, and support Prometheus metrics and OpenTelemetry tracing.
Who it is for. For financial analysts and developers who want standalone MOEX market and portfolio risk MCP servers rather than the whole demo web app.
Good fit when
- You need VaR, correlations or peer comparison for a MOEX stock portfolio
- You need a CFO-oriented liquidity report without manual calculations
- You are ready to deploy the moex_iss_mcp and risk_analytics_mcp Docker images separately
Not a fit when
- You only need simple quote access without risk analytics: a lighter MOEX MCP server is simpler
- You do not want to navigate a large demo repository structure to reach two nested servers
Example request
Calculate VaR for a portfolio of SBER, GAZP and LKOH and suggest a rebalanceLimitations
0 stars, a single author, 321 files in the repository, of which the MCP servers are only part of a larger multi-agent demo built on FastAPI, Next.js and OpenRouter. The README's quick start is oriented around running the whole stack via make local-up; to run one server in isolation you need to build its specific Dockerfile manually.
How to disable. Remove the moex-iss and risk-analytics blocks from your MCP client config and stop the corresponding Docker containers.
MCP
- Transport
- http
- Authentication
- not required
| Environment variables | |
|---|---|
| MOEX_ISS_BASE_URL | ISS base URL, defaults to https://iss.moex.com/iss |
| MOEX_ISS_RATE_LIMIT_RPS | Requests-per-second limit against ISS |
| MOEX_API_KEY secret | Needed only for paid ISS access |
| RISK_MAX_PORTFOLIO_TICKERS | Max tickers in a portfolio risk calculation, default 50 |
Security check
- Only reads market data and computes analytics, never places orders
- Does not connect to brokerage accounts
README in short
The main README describes a multi-agent solution with an A2A agent, market_data and risk_analytics subagents, a Next.js dashboard and an SSE stream. Inside, moex_iss_mcp/README.md and risk_analytics_mcp/README.md detail each server's tools, all environment variables, a local run via uv, building and publishing linux/amd64 Docker images, and a troubleshooting section for ISS limits and timeouts.
FAQ
Do I have to run the whole multi-agent stack?
No, moex_iss_mcp and risk_analytics_mcp can be built and run as separate Docker images from their own Dockerfiles, without the agent or web UI.
Do I need a MOEX key for risk analytics?
Usually not; MOEX_API_KEY is listed as a variable needed only for paid ISS access.
Related
166 skills for scientific work: bioinformatics, cheminformatics, clinical data, geospatial analysis and 100+ databases
Google's open-source MCP server for databases: ready tools for Postgres, MySQL, BigQuery, Spanner and more, plus custom tools in tools.yaml
Official Hugging Face skills: Hub operations via the hf CLI, datasets, model training, Spaces, evals and deployment
A visualization language for agents and an MCP server: neat charts from a simple semantic spec