Token Metrics

Fetches current and historical cryptocurrency price data, market capitalization, trading volumes, and market metrics through standardized MCP tool interface (get_tokens_price, get_tokens_data, get_market_metrics). The system acts as a middleware layer translating MCP tool calls into authenticated HTTP requests to the Token Metrics API, caching responses to reduce latency and API quota consumption. Supports batch queries for multiple tokens and configurable time windows.

Generates actionable trading signals (buy/sell/hold recommendations) and grades trader performance using Token Metrics' proprietary algorithms through get_tokens_trading_signal and get_trader_grade tools. The system wraps Token Metrics' signal generation engine, returning structured recommendations with confidence scores and historical accuracy metrics. Signals are computed server-side and delivered as JSON payloads containing signal type, strength, and supporting rationale.

Implements flexible API key authentication supporting both environment variables (for CLI/local deployment) and HTTP headers (for HTTP/OpenAI transports). The system validates API keys at server startup for CLI mode and on each request for HTTP modes, returning 401 Unauthorized if key is missing or invalid. Authentication is decoupled from tool implementations, allowing tools to assume authenticated context.

Provides production-ready Docker images and Kubernetes manifests for deploying Token Metrics MCP server at scale. The system includes multi-stage Dockerfile for optimized image size, Kubernetes deployment/service/ingress manifests for orchestration, and CI/CD pipeline (GitHub Actions) for automated testing and image publishing. Deployment supports environment variable configuration, health checks, and resource limits.

Implements HTTP Server-Sent Events (SSE) transport for streaming responses from long-running tool operations (scenario analysis, report generation). The system uses HTTP/SSE protocol to send partial results and progress updates to clients in real-time, avoiding request timeouts for expensive computations. Clients receive streaming JSON objects that can be processed incrementally as they arrive.

Implements OpenAI-compatible HTTP server that exposes Token Metrics tools as OpenAI function calling schemas. The system translates MCP tool definitions into OpenAI function calling format, handles OpenAI-specific request/response serialization, and manages function call execution within OpenAI's function calling workflow. Allows OpenAI API clients to call Token Metrics tools directly without MCP client implementation.

Computes technical analysis indicators including resistance/support levels, price correlation between tokens, and momentum metrics through get_tokens_resistance_and_support and get_tokens_correlation tools. The system queries Token Metrics' technical analysis engine which performs statistical analysis on historical price data to identify key price levels and cross-token relationships. Results are returned as structured JSON containing price levels, confidence intervals, and correlation coefficients.

Performs scenario-based analysis and computes advanced quantitative metrics (Sharpe ratio, volatility, Value-at-Risk) through get_tokens_scenario_analysis and get_tokens_quant_metrics tools. The system executes server-side Monte Carlo simulations and statistical calculations on historical token data to project potential outcomes under different market conditions. Results include probability distributions, risk metrics, and performance projections returned as structured JSON.

Aggregates and analyzes market sentiment from social media, news, and on-chain data through get_sentiment and related tools. The system collects sentiment signals from multiple sources (Twitter/X, Reddit, news feeds, blockchain metrics) and computes aggregate sentiment scores using natural language processing and statistical aggregation. Results include sentiment polarity scores, trend direction, and source-specific breakdowns returned as JSON.

Generates comprehensive AI-powered research reports on cryptocurrencies and market trends through get_tokens_ai_report tool. The system uses Token Metrics' LLM-based analysis engine to synthesize market data, technical analysis, sentiment, and fundamental metrics into narrative research reports. Reports include executive summaries, risk assessments, and investment theses returned as structured text with embedded data references.

Tracks cryptocurrency investors, funds, and their portfolio holdings through get_crypto_investors tool. The system maintains a database of known crypto investors and funds, exposing their portfolio compositions, historical performance, and investment patterns. Results include investor profiles, holdings lists, and performance metrics returned as structured JSON.

Provides access to cryptocurrency indices and portfolio performance metrics through get_indices and get_indices_performance tools. The system computes weighted indices across token baskets (market cap weighted, equal weighted, custom) and tracks portfolio performance against benchmarks. Results include index values, constituent weights, and performance attribution returned as JSON.

Implements three distinct server transport modes (stdio CLI, HTTP/SSE, OpenAI-specific HTTP) from single codebase, allowing the same tool ecosystem to serve local development, web applications, and OpenAI integrations. The system uses MCP protocol's standardized tool schema to define tools once and expose them across all transports without code duplication. Each transport mode handles authentication, request routing, and response serialization independently while sharing core tool implementations.

Defines 21+ cryptocurrency analysis tools using MCP's standardized tool schema, providing consistent parameter validation, error handling, and response formatting across all tools. The system uses JSON schema to define tool inputs (required/optional parameters, types, constraints) and outputs (response structure, data types). Tool definitions are validated at server startup and used to generate OpenAI function calling schemas, HTTP endpoint documentation, and CLI help text automatically.