Discover the Best AI Tools

A FastMCP server for hybrid semantic search and reranking over philosophy and humanities texts, with author and language filters. Self-hosting requires a populated Qdrant collection; installing the server alone does not supply the corpus.

An MCP client for searching Orbit Sentinel’s space regulatory filings and entity profiles, including material from FCC, ITU, UNOOSA and FAA-AST. Local stdio and hosted HTTP access use the remote service. Access requires credentials; the MIT client does not include the hosted database or guarantee complete regulatory coverage.

An MCP interface to Octodamus market data, sentiment summaries and model-derived trading signals for crypto, equities and prediction markets. It offers free and authenticated or paid endpoints through a hosted API. Signal confidence, estimated probabilities and example track records are provider outputs, not independently verified accuracy or a guarantee of investment returns.

MCP client for ForgeMesh’s hosted sequence-anomaly analysis across blockchain, aviation and GitHub data. Paid tools use USDC on Base via x402; by default the adapter returns payment challenges rather than settling them. Some scans are best-effort and report source-coverage limits.

MCP client for paid Kronos market-context ranges, risk information and decision audit records from hosted CoinOpAI endpoints. A configured wallet automatically signs per-call USDC payments on Base via x402. Forecasts can be wrong; this is not a buy/sell oracle.

Research environment for LLM trading agents with point-in-time masking, entity and date anonymization, deterministic replay and training-trajectory export. Focuses on Chinese A-shares; has no live broker interface or market-impact model, and source-data terms apply.

Python environment for training and backtesting reinforcement-learning or rule-based trading agents using tick or OHLC data. The project is work in progress; built-in training examples and live broker integration are not complete.

Open Data Platform for integrating public, licensed and proprietary data into Python workflows, research dashboards and AI applications through APIs and MCP; the code license does not grant access to separately licensed data or services.

Python random-sequence generator with data calibration and presets for fat tails and volatility clustering, plus reproducible descriptive benchmarks; it models statistical properties and does not forecast future prices.

Python framework for building, training and evaluating reinforcement-learning trading agents with composable environments, action schemes, rewards and data feeds; current setup requires Python 3.12+ and training dependencies, and results depend on trading costs.

HTTP and MCP data service with a discoverable catalog and individually priced calls; supported purchases use x402 and USDC on Base and require a payment-capable client, wallet authorization and sufficient funds.

MCP server for stock and cryptocurrency data, locally computed technical indicators and portfolio analysis using Yahoo Finance and CoinGecko; upstream limits apply, and hosted plans have separate quotas and pricing.

Rust-based technical indicator library for streaming and batch calculations, with Python, Node.js and WebAssembly bindings and a shared C interface; it supplies calculations rather than predictions or trading advice.

PyTorch research stack for factor computation, model training, portfolio construction and cost-aware backtesting, with synthetic demos and report exports; it is not production-ready and backtests do not establish future returns.

Modular quantitative-trading framework connecting data processing, portfolio-weight strategies, backtesting and Alpaca paper or live execution; brokerage execution requires account credentials and does not guarantee returns.

Python client and MCP server for retrieving Asian company financial reports or selected sections as Markdown from the DataSinking service; an API key is required and document quotas depend on the service plan.

Independent MCP server built on sec-edgar-toolkit for company filings, financial statements and insider transaction disclosures, returning filing links; it requires a configured SEC request identity and includes no HTTP authentication.

Self-hosted MCP server for stock analysis, screening and portfolio tracking using yfinance, with optional backtesting and model-assisted research extras; research requires separate model and search-service access.

Development-stage MCP adapter for the Interactive Brokers Web API, using Docker services for a Client Portal Gateway and MCP routing; it requires an authenticated brokerage session and can expose trading operations.

Local MCP server and CLI connecting assistants to OKX market data, account and trading operations, with read-only and module-filtering options; trading functions use exchange credentials and can place real orders.

Experimental command-line client with a built-in MCP server for Kraken market data, account operations and paper or live trading; authenticated live commands can execute real financial transactions.

MCP interface to Alpaca market data, accounts and order operations, requiring account API credentials; paper trading is the default, while explicitly configured live mode can execute real trades.

MCP connector for tennis scores, players and fixtures, with additional odds, statistics and model estimates depending on the data-service plan; an API key is required.

Independent MCP server for searching and retrieving FRED economic time-series data using a FRED API key; it is not affiliated with the Federal Reserve.