TREMOR vs
SimpleFunctions.

SimpleFunctions's thesis system takes a plain-language sentence and decomposes it into a causal tree of testable sub-claims. Each sub-claim gets a probability; SF scans Kalshi and Polymarket for tradeable edges that match, then runs an evaluation heartbeat — news scan, price refresh, milestone check, LLM evaluation, confidence update. You can inject external signals via /api/thesis/{id}/signal and fork public theses. TREMOR's AI assistant is a conversational query layer over market data — useful for exploring the dataset, not for autonomous thesis tracking or trading-edge discovery.

No. SimpleFunctions does not expose a browser-based SQL interface. SF exposes a REST API, a 56-tool MCP server, and a CLI with 60+ commands for programmatic access. If you need to write ad-hoc SQL across the full market dataset interactively in a browser, TREMOR's SQL terminal is the better fit. If you need agents, automated pipelines, or Claude Code integrations to consume prediction market data, SF's REST and MCP surfaces are purpose-built for that workflow.

Portfolio Autopilot is SF's autonomous trading agent. It runs a 1M-context LLM against 13 data sources — cross-venue prices, orderbook depth, computed indicators, thesis signals, regime labels, and more — and passes every candidate action through a 7-gate risk cascade before execution: kill switch, position limits, drawdown gate, regime check, and additional guards. It is not a recommendation engine; it executes autonomously subject to the configured risk parameters. No comparable autonomous execution layer is available in TREMOR.

Based on publicly available information, TREMOR does not publish an MCP server. SimpleFunctions exposes 56 tools via one command: claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp. Any MCP-compatible client — Claude Code, Cursor, or a custom agent — can then call SF endpoints natively, including market search, thesis creation, orderbook queries, calibration data retrieval, and cross-venue arbitrage scanning, without custom integration code.

SF pre-computes six indicators across 48K+ active contracts: IY (implied yield — annualised return assuming YES resolution), CRI (cliff risk index — price sensitivity to a binary reversal near settlement), LAS (liquidity availability score), EE (event overround — sum of all YES prices minus 1, measuring market-maker edge), τ-days (time to settlement), and regime label (adverse-selection classification). All are accessible at /screen or via the MCP server. TREMOR provides raw price and volume data; custom indicator computation is left to the user.

Hit /api/calibration with no authentication required. The endpoint returns SF's own Brier scores broken down by venue, category, and price bucket — computed on the past 90 days of resolved contracts. Current figures: Kalshi 0.20, Polymarket 0.12 on T-24h price. A lower Brier score is better (0 = perfect). TREMOR does not publish comparable calibration data. Publishing your own error metrics publicly means accuracy claims are checkable rather than asserted.

The discrepancy reflects different definitions of coverage. TREMOR counts all markets it indexes, including recently closed and low-liquidity markets. SF's 48K+ figure reflects contracts that are currently open, have non-zero liquidity, and pass an activity threshold necessary to compute the indicator suite meaningfully. Neither number is wrong; they measure different things. If breadth across historical and low-liquidity markets matters for SQL exploration, TREMOR's 140K+ is the relevant figure. If you need live, liquid contracts with pre-computed indicators, SF's 48K+ is the relevant universe.