Marketlens vs
SimpleFunctions.

SimpleFunctions exposes Polymarket orderbook depth at /api/public/market/{ticker}?depth=true and normalised historical prices across both Kalshi and Polymarket, with daily exports on HuggingFace and Kaggle. Marketlens specialises in tick-level orderbook replay via its Python SDK — if your workflow requires replaying individual tick sequences through a Polymarket-native backtest harness, Marketlens has invested specifically in that surface. SimpleFunctions focuses on cross-venue normalisation, computed indicators, and agent-readable outputs rather than tick-replay infrastructure.

POST /api/thesis/create takes a plain-language claim and decomposes it into a causal tree of testable sub-claims. The system scans Kalshi and Polymarket for contracts that price each node, propagates implied probabilities up the tree, and runs an evaluation heartbeat: news scan → price refresh → milestone check → LLM evaluation → confidence update. You can inject new signals at any time via /api/thesis/{id}/signal, and public theses are forkable. No current prediction market data product offers comparable thesis decomposition with automatic contract mapping and evaluation cycles.

Backtesting — Marketlens's focus — replays historical data offline to evaluate a strategy's past performance. Portfolio Autopilot is a live execution system: a 1M-context LLM ingests 13 data sources in real time, then passes a 7-gate risk cascade (kill switch, position limits, drawdown gate, regime check, and additional controls) before placing any trade. The two address different phases of a trading workflow. Marketlens is the right tool for historical validation; SF Autopilot handles live autonomous execution with layered risk controls.

Based on public documentation, Marketlens is Polymarket-only and does not index Kalshi. If your research or agent pipeline needs unified coverage across both venues — normalised prices, cross-venue matched pairs, or arbitrage opportunity detection — SimpleFunctions indexes 48K+ active contracts from Kalshi and Polymarket on the same feed, accessible at /api/public/markets and /api/public/cross-venue/pairs.

Yes. SimpleFunctions exposes a language-agnostic REST API that any Python script can call using the standard requests library — no dedicated SDK is required. There is also an npm CLI with 60+ commands and a 56-tool MCP server for agent frameworks. Reads require no authentication. Marketlens provides a dedicated Python SDK with Polymarket-native conventions that may reduce boilerplate if your stack is already Python-only and Polymarket-only and you need tick-replay features.

Run claude mcp add simplefunctions --url https://simplefunctions.dev/api/mcp/mcp in your terminal. That registers all 56 tools — market search, thesis creation, indicator lookup, orderbook depth, world snapshot, cross-venue pair scan, and more — directly inside the MCP client. No API key is required for read-only tools. Marketlens does not publish an MCP server, so it is not directly accessible from within Claude Code or Cursor via the MCP protocol.

/api/calibration returns SimpleFunctions' own Brier scores broken down by venue, category, and price bucket, computed from the past 90 days of resolved contracts. Published figures: Kalshi 0.20, Polymarket 0.12 on T-24h price. Most data products assert their probabilities are reliable; SF lets you verify that numerically before trusting its outputs in an agent pipeline. This is particularly relevant when you are weighting inputs from multiple data sources and need a defensible accuracy baseline rather than a marketing claim.

Yes. If your primary need is tick-level Polymarket orderbook replay for offline backtesting — especially within a Python-native quant workflow — Marketlens is purpose-built for that use case and has rate-limit tiers (up to 1,800 req/min on Enterprise) calibrated for high-frequency data access. SimpleFunctions is designed around cross-venue agent surfaces, thesis automation, and live execution rather than a historical tick-replay harness. The two products serve different workflows and are not direct substitutes.