About

About MarketIntell

We built the synthesis layer for market decisions. One natural-language question, 27+ external services queried in parallel, multi-tier reasoning, and a sourced trade setup with a public 24-hour outcome stamp on every signal.

Mission

What we're for

Every market decision today — retail trader sizing a position, fund analyst evaluating a setup, autonomous AI agent rebalancing — runs into the same wall: the data is everywhere, the reasoning is nowhere. MarketIntell exists to be the single brain that synthesizes the data, explains the call, and publishes the outcome.

We sell intelligence, not access. Pricing reflects compute + reasoning quality; the data underneath is racing to free.

Team

Who builds this

Founder & engineer

Solo founder

Full-stack engineer with multi-year crypto trading experience. MarketIntell was built out of personal scars from deciding on incomplete data — the analyst desk we wished we had. Solo founder; engineering hire #2 in progress.

Hiring engineer #2

Open role

Full-stack TypeScript, comfortable with Bun + Next.js + Prisma, LLM-systems experience. Equity-heavy compensation. If you read the codebase first then send a 5-line note about the most interesting bug you'd fix, we will reply.

[email protected]

Story

How we got here

MarketIntell started as a personal trading tool — a way to get the multi-source synthesis no commercial product ships. Within three months it had a full specialist desk, dozens of live data providers, and an autonomous Polymarket bot calling the same API for hundreds of trades with no human in the loop.

By Q3 2026 the product had 3 asset domains (crypto, US equities, India), a 17-tool MCP server live in Claude Desktop and Cursor, and the public 30-day accuracy widget that no competitor publishes a denominator for.

See /updates for the monthly audit log.

Product

The shape of MarketIntell today

  • 3 asset domains: crypto is the deep, production desk; US equities (with options) and India equities are real but newer desks (Beta). Macro and prediction-market data are context layers the desks read, not standalone asset classes.
  • 4 reasoning tiers: T1 fast agent, T2 deterministic CoT, T3 supervisor, T4 autonomous agents.
  • 27+ external services queried in parallel, all behind one tool surface.
  • 17 MCP tools live at api.marketintell.ai/mcp/sse.
  • Public 30-day rolling win rate at /track-record; per-signal post-mortems shipping next.
  • Closed cost loop (per-tenant monthly_cost_cap_cents) and closed eval loop (Claude-as-judge sampler + golden bench in CI).
  • Tested, CI-gated, and shipped under real release tags — production software, not a demo.

Contact

Where to reach us

Mailing address available on request to enterprise customers under NDA. We are a distributed team operating across multiple jurisdictions; entity disclosure on commercial engagements only.

Don't take our word for it. Open a free account, ask a question, and audit the trading wallet yourself.

Or get the monthly audit log at /updates.