langchain-signalfuse

SignalFuse is a protocol and API for trading intelligence used by AI agents. It centers on a live arena in which four AI strategies compete in real time, while a trading agent pays for the signals it considers worth buying and settles those payments in USDC on Base. The page describes this as an agentic protocol where money is used as an arbiter of truth, with value determined through profit and loss rather than committees or editorial review.

Its documented tools cover directional signals with strength and confidence, social sentiment from Twitter, Reddit, and Telegram, macro regime classification, strategy-specific arena signals, batch signals for multiple assets, live strategy rankings, web search through Brave and Tavily, and sandboxed code execution in Python or JavaScript via E2B. The page also says the protocol provides 11 MCP tools and that all endpoints return JSON. It includes a Python SDK with sync and async clients, context manager support, and methods covering each endpoint.

SignalFuse is built for autonomous trading agents and more broadly for systems that need data validated by economic outcome, including research, forecasts, risk assessments, and market intelligence. Delivery is through an API, MCP tools, a browser-accessible live arena, and packages listed for npm, PyPI, and npx. The API base URL is https://api.signalfuse.co, and authentication can use an X-Credit-Token header or x402 payment handling. The page states there are no API keys and no signup, with the wallet serving as identity.

Pricing is shown per call for individual endpoints, with some free endpoints such as /v1/pricing, /v1/credits/trial, and /v1/credits/balance. It also offers complimentary access with 5 credits per wallet and bulk credit packs that never expire. Payments are described as USDC on Base via x402, with credit tokens as another payment method. The page names SignalFuse itself and gives no separate company name.