Converge vs
SimpleFunctions.

SF's causal thesis system takes a plain-English sentence via POST /api/thesis/create and decomposes it into a causal tree of testable sub-claims. Each node gets a probability that propagates up the tree. An evaluation heartbeat runs on a recurring cycle — news scan, price refresh, milestone check, LLM evaluation, confidence update — and you can inject signals at any node via /api/thesis/{id}/signal. Public theses are forkable. Converge is a trading terminal; it has no thesis decomposition or evaluation infrastructure.

Based on publicly observable information at converge.bet as of 2026-04, Converge presents as a trading terminal UI rather than a developer API product. No public REST endpoint documentation was found. If you need a programmatic interface with a documented REST spec, SF publishes openapi.json, llms.txt, and a 56-tool MCP server, all without requiring authentication for read operations.

SF's current indexed venues are Kalshi and Polymarket, covering 48K+ active contracts. Converge's stated coverage includes Limitless as a third venue alongside Polymarket and Kalshi. If Limitless-specific data or execution is a hard requirement for your workflow, Converge's terminal addresses that directly.

Portfolio Autopilot is SF's autonomous trading agent: a 1M-context LLM with access to 13 data sources and a 7-gate risk cascade — kill switch, position limits, drawdown gate, regime check, and others — before any order is placed. It reads the thesis system, screens contracts for edges, and executes only when all gates clear. No current prediction market terminal, including Converge, publishes an equivalent autonomous agent layer.

Yes. Run claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp to install 56 tools covering price queries, thesis creation, signal injection, orderbook depth, cross-venue pairs, and more. The MCP server requires no API key for read tools. Converge has no published MCP server, so there is no equivalent one-line integration path for agent frameworks.

GET /api/calibration returns SF's own Brier scores segmented by venue (Kalshi and Polymarket), market category, and price bucket, computed over the past 90 days using T-24h prices. Current figures are Kalshi 0.20 and Polymarket 0.12. This lets you audit SF's forecast accuracy before relying on it in a model or agent pipeline. Converge does not publish calibration metrics; its product is execution, not forecast accuracy tracking.

They solve different problems. Converge is a custody-free terminal built for human traders who want to execute across Polymarket, Kalshi, and Limitless in a single interface with arbitrage detection. SF is an API, CLI, and MCP server designed for agents, researchers, and developers who need calibrated data, computed indicators, causal thesis modelling, and autonomous execution infrastructure — not a human-facing trading UI.