Markets
Lenz Relaunched a Multi-Model Fact-Checking API on the Claim That Frontier Models Disagree on 23% of Real Claims
Lenz launched again on Product Hunt on 2026-08-27 at #10 with 93 points, extracting verifiable claims from text and validating them through independent source search, multi-model debate and a review panel, returning scored verdicts with full source attribution. The founders' stated premise is that on 23% of real-world claims frontier models disagree significantly, which is why they refuse to rely on any single model. It ships as both an API and an MCP server built with GStack, Claude Code and the Perplexity API, and it is the second launch after an April 3, 2026 debut, so treat the traction claims as unverified.
Source
↳ Follow the thread