The fact-checker for AI

Confidence
isn't truth.

Your model always sounds sure. Mephistopheles pulls apart every claim it makes, grounds each against independent sources and tools, and tells you what's true, what's wrong, and what it can't verify — with citations.

Verifying an answerrunning…
Mount Everest is the tallest mountain on Earth at 8,849 m, first summited in 1953 by Hillary and Tenzing — located in the Andes.
supported
Everest is the highest mountain above sea level (8,849 m).
↳ britannica.com · nationalgeographic.com
supported
First summited in 1953 by Hillary and Tenzing Norgay.
↳ rgs.org · himalayandatabase.com
contradicted
Everest is located in the Andes.
↳ 3 sources: it is in the Himalayas, not the Andes.
hallucination risk
high
How it works

It doesn't trust the answer. It checks it.

01

Break it into claims

Every factual statement is pulled out as a standalone, checkable claim — the opinions and fluff left aside.

02

Ground each one

Each claim is checked against several independent sources, exact tools, and your own trusted documents — not the model's memory.

03

Return a verdict

Supported, contradicted, or unverifiable — each with citations and a confidence, plus one honest risk score for the whole answer.

Why it's different

Grounds what's checkable. Flags what isn't. Never guesses.

  • Multiple independent sources. A claim is only marked wrong when independent evidence agrees — not on one stale hit.
  • Exact tools for exact facts. Math and conversions go to a calculator, right every time — no search needed.
  • Your own knowledge. Point it at your documents and they become the authority for your domain.
  • Honest about limits. When it can't verify something, it says so — instead of inventing confidence.
One line to add it

Works with the AI you already use.

Point your OpenAI base_url at Mephistopheles and every answer comes back verified — or call /verify with any model's output. Nothing else in your code changes.

# same code, one new line
client = OpenAI(
  base_url = "https://mephistopheles.ai/v1",
  api_key  = "mk-your-key",
)
# → replies now carry a per-claim
#   verdict, citations & risk score

When the facts matter,
don't trust — verify.

Free to start. One line to add. Every claim accounted for.

Get an API key