Catch AI hallucinations before they ship.

Lenz is a fact-checking API for model-generated text. It checks factual statements against independent public sources, not against the model that produced them, and returns structured results.

Hallucination detection against public sources.

Send Lenz an answer your model produced. It pulls out the factual statements in it and checks each against independent public sources, using models from multiple vendors. For a full verification, models argue for and against each statement on the evidence, three automated reviewers assess it, and the result ships with the trace.

A model grading its own output is not an outside check. Lenz sits outside the model vendor and outside your own team, and the evidence it cites is public, so a customer or an auditor can follow it.

The API: two different checks, one sequence.

Lenz supplies the verdicts. Your gate applies your thresholds.

1. 1

   An answer is drafted

   A support reply, a RAG answer, a summary — anything a model wrote that a customer is about to read.

2. 2

   Extract the statements

   POST /extract splits the answer into single factual statements, up to 20, each self-contained enough to check on its own. Opinions, recommendations, predictions and questions are left out. It costs no credits, capped at 1,000 calls a day per account.

3. 3

   Sort them, then establish the ones that matter

   POST /assess is a judgement: a panel of models returns a verdict and a confidence level per statement, synchronously in about 10 seconds, up to 20 statements a call — no score and no sources. Send the negative and low-confidence ones to POST /verify, an investigation that searches independent public sources, has models argue both sides and three automated reviewers assess the arguments — about 90 seconds a statement, asynchronous, with a webhook when it lands.

4. ✓

   Your gate decides

   Your thresholds and your routing: what clears them ships, and what does not goes to your reviewer with the verdict, the sources and the reasoning already attached. Lenz returns results — it never holds, edits or publishes your text.

5. 5

   Ship, with the record

   A full verification keeps its trace: the statement as framed, every source with its date, the arguments on both sides and the reasoning that weighed them. Submissions through the API are private by default; visibility: "unlisted" returns a link you can pass to whoever asks.

The sequence is extract → assess in bulk → verify the uncertain results → ask follow-up questions about a verification. Assess to sort; verify to establish.

Fact-checking RAG answers.

Retrieval narrows what the model says; it does not make it true. Lenz checks the factual statements in the generated answer against public sources. It does not compare the answer with your retrieved documents: faithfulness to your own context is a different check, and not this one.

Check LLM output with Python.

# model output in — the statements your policy holds back, out
from lenz_io import Lenz

client = Lenz(api_key="lenz_...")
extracted = client.extract(text=model_output)
# identified_claims is empty when the text yields a single statement
claims = extracted.identified_claims or [extracted.claim]
r = client.assess(claims=claims)   # one row per statement, same order

# your routing policy. "Error" means the item was not checked — never ship it as clean
for c in r.claims:
    if c.verdict in ("False", "Mostly False", "Mixed", "Error") or c.confidence == "low":
        print("HOLD:", c.claim, c.verdict, c.confidence)
# HOLD: Python 3.12 runs about 20% faster than Python 3.11 ... False high

What comes back.

Selected fields from the recorded responses for the example above, which Lenz wrote and published. Your API submissions are private to your account.

{
  "claim": "Python 3.12 runs about 20% faster than Python 3.11 on the official Python benchmark suite.",
  "language": "en",
  "verdict": "False",
  "confidence": "high",
  "verification_url": "https://lenz.io/api/v1/verifications/b5968a41"
}

{
  "status": "completed",
  "result": {
    "verification_id": "b5968a41",
    "claim": "Python 3.12 runs about 20% faster than Python 3.11 on the official Python benchmark suite.",
    "depth": "standard",
    "verdict": "False",
    "confidence": "high",
    "lenz_score": 1,
    "key_finding": "Python 3.12 averages roughly 4–5% faster than Python 3.11.",
    "sources": [
      { "url": "https://docs.python.org/3/whatsnew/3.12.html" }
    ]
  }
}

A full verification returns the fields your code branches on — verdict, confidence, lenz_score, key_finding and sources — so the routing decision is yours to make in code: publish, hold, or send it to your reviewers.

Lenz checks factual statements in text against public sources.

Frequently asked questions