Markium vs
SimpleFunctions.

Markium is a Polymarket-focused analytics platform that provides wallet-level tracking, leaderboard rankings, and watchlist alerting. It is designed for traders who want to monitor on-chain wallet behaviour, follow which addresses are outperforming, and receive notifications when watched contracts move. It is not primarily an API product or agent-integration layer — its audience is Polymarket participants who want a data-enriched dashboard around social signals and wallet activity.

No. SimpleFunctions does not expose a dedicated wallet-attribution or leaderboard surface — that is Markium's specialisation. SF's comparative strengths are the agent-integration layer: the 56-tool MCP server, the causal-tree thesis system, Portfolio Autopilot, and computed indicators across 48K+ contracts. If wallet tracking and leaderboard monitoring are your primary workflow, Markium addresses that directly. If you need autonomous trading, thesis evaluation, or agent-first market intelligence, SF is the more relevant stack.

POST /api/thesis/create accepts a plain-language sentence and decomposes it into a causal tree of testable sub-claims. Each node receives a probability assignment, and SF scans Kalshi and Polymarket for contracts that map to those sub-claims, surfacing tradeable edges. An evaluation heartbeat then runs on each thesis: news scan, price refresh, milestone check, LLM evaluation, and confidence update. External signals can be injected at /api/thesis/{id}/signal, and any public thesis is forkable. No current prediction-market analytics product offers this pipeline.

Portfolio Autopilot is SimpleFunctions' autonomous trading agent. It uses a 1M-context LLM that ingests 13 data sources, then passes proposed trades through a 7-gate risk cascade — including a kill switch, position limits, drawdown gate, and regime check — before any execution occurs. Positions are grounded in the thesis system's causal model rather than raw price momentum alone. This autonomous execution layer is not offered by any current Polymarket analytics or wallet-tracking product.

Yes. Run claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp to expose all 56 tools to your MCP client. From that point, your AI assistant can query live market prices, fetch cross-venue arbitrage pairs, retrieve computed indicators, create and evaluate theses, and interact with the full SF data model — without writing HTTP calls manually. The server conforms to the MCP specification and works with Claude Code, Cursor, and any other MCP-compatible client.

GET /api/calibration returns SF's own Brier scores broken down by venue (Kalshi and Polymarket), category, and price bucket, computed over the past 90 days. On T-24h prices, SF scores Kalshi at 0.20 and Polymarket at 0.12. This is a live endpoint — you can curl it yourself to verify the numbers on this page. Publishing self-audited accuracy metrics publicly is uncommon among prediction-market analytics products; most either omit calibration data or describe it only in qualitative terms.

Both products surface Polymarket data, but with different emphasis. Markium wraps it in a wallet-analytics and alerting dashboard suited for traders who follow positions and leaderboards. SimpleFunctions wraps it in an agent-integration layer — computed indicators, cross-venue pricing, thesis evaluation, and an MCP server. If you need a human-facing monitoring dashboard with social signals, Markium fits. If you need a programmable, agent-consumable market intelligence layer, SF fits.