Capability

Every answer traces back to a passage

Explainability isn't a heatmap or a confidence score — it's the exact sentence the model relied on. Open any claim and you see the source passage, the page, and the document it came from.

Source-level provenance, not feature-level

The unit of explanation is a quoted passage from a real document. That's the level of detail a compliance officer, an auditor, or a reviewer actually needs — anything more abstract gets ignored.

Reviewable in one click

Every citation opens the source passage. The reader verifies the supporting text without changing tools or losing their place in the answer.

Works for teams, not just users

When a colleague picks up an answer the next day, the same evidence is attached. Explanations don't live in one analyst's head — they live on the answer.

Explainability FAQ

Is this the same as a confidence score?
No. Confidence scores tell you how sure the model is. Explanations tell you what the model relied on. The second is more useful when an auditor asks.
What if the source is wrong?
The model quotes what's in the source. If the source itself is wrong, the answer reflects that — and you can see exactly which document caused it.
What does it cost?
Source-level explanations are part of every plan.
How do I get started?
Sign in and ask a question — the citations are clickable from the first answer.

Open every claim and see the source

Sign in to try it on your documents.

Related