PreReason

About PreReason

PreReason is a market context API designed to be consumed by AI agents rather than by dashboards, and the design follows from that premise more rigorously than most 'AI-ready' data products. It pulls Bitcoin on-chain metrics, macro indicators, mining and liquidity data from six-plus sources and returns them pre-reasoned: trend direction as rising, falling or flat, 7/30/90-day momentum, percentile rankings, cross-asset correlations, regime classification and confidence scores, all in one structured response. The alternative an agent developer otherwise faces is integrating five data sources, parsing raw numbers and building an interpretation layer, which is where most agent projects lose their time. The delivery detail worth noting is the LLM-native markdown format: responses are formatted for how language models read rather than as JSON, which the vendor measures at roughly 60% fewer tokens for the same content. Given that agents pay per token on every call, that is a direct cost reduction, not a formatting preference. Distribution is agent-first throughout — REST API, an MCP server installable with npx @prereason/mcp, and llms.txt — working with Claude Desktop, Claude Code, Cursor and Windsurf. Eighteen market briefings are split six free, six on Basic and six on Pro, with the free tier allowing 60 requests an hour and 500 a day against 30 days of rolling data, no card required.

Key Features

PreReason Pros & Cons

Tags