Amalgix

Pulls evidence-backed findings from financial filings and contracts using a two-stage pipeline: one model extracts claims with source references, a second model verifies them against the original text. You get structured output with confidence scores and verification status for every finding. Handles SEC 10-K/10-Q analysis by ticker or CIK, contract risk review for clauses and obligations, and general document analysis. Runs on x402 pay-per-call over streamable HTTP, settling in USDC on Base or Solana. Cost runs 60-75% below running two passes of Opus or GPT yourself because it routes to cheaper models at each stage. Eight tools total, including free cost estimation before you commit to a paid analysis run.