groundlens

10/14/2026

GroundLens checks an LLM answer against its sources without a model, decides PASS, REVIEW or FAIL under your policy, and keeps a signed record.

• Exact checks first. Every figure must be in the passages or follow from figures that are; a real figure from another company, contract or year fails; the answer must give the kind and unit the question asked for.
• A policy you own turns the evidence into a decision, with reasons.
• A signed, hash-chained record of every decision, stored before the decision returns and verifiable offline.
• Your LLM judge, if you want one, only sees the answers the exact checks could not reject.

pip install groundlens

import groundlens as gl

verifier = gl.Verifier(sink=gl.sinks.jsonl("records/"))
record = verifier.check(answer="Revenue was €42.1 million.", passages=["Revenue reached €41.2 million."])
print(record.decision)   # FAIL

Works with Langfuse, Grafana and any OpenTelemetry backend, and MCP clients. Rust engine, Python package, Apache-2.0.

Home · GitHub repository · Documentation · PyPI