Fairness you can verify
A hiring tool that claims to be fair should let you check. Open source means the composite math, the scorecard weighting, and the audit chain will all be readable, not a black box you have to trust.
You can’t ask teams to trust a fairness claim they can’t inspect. FairHire is built to be fully open source under the MIT license. Read the code, run it yourself, and see exactly how every hiring decision is reached. The public repository is on its way.
A hiring tool that claims to be fair should let you check. Open source means the composite math, the scorecard weighting, and the audit chain will all be readable, not a black box you have to trust.
Self-host the Community edition and every candidate record, note, and decision lives on your infrastructure. No per-seat data hostage situation, no export fees.
Because the hosted product is the same MIT-licensed codebase, you can move from hosted to self-hosted (or back) whenever you want. The exit is always open.
Fork it, add a stage, wire a new scorecard type, or expose a custom MCP tool. FairHire is built on composable blocks so the hiring last-mile is yours to shape.
FairHire assembles reusable building blocks (auth & multi-tenant orgs, an MCP agent spine, forms, collaboration, reporting, and an immutable audit log) and adds only the hiring-specific logic on top: the funnel, weighted scorecards, paid trials, and the fairness trail. That’s what keeps the codebase small enough to actually read.
See the agent-native platformThe full platform runs hosted today: funnel, weighted scorecards, paid trials, the fairness trail, and the MCP agent surface.
The source goes public under the MIT license: the composite math, the scorecard weighting, and the hash-chained audit chain, all readable.
A documented self-host path: one service, one container, your infrastructure, your data. Move between hosted and self-hosted freely.
The repository is coming. Until then, run the hosted beta and read the docs to see exactly how the fairness math works.