The same checks, in under half a minute.

Lite is a smaller panel for products that need an answer while the user waits. It keeps every evidence check that decides what reaches the brief and spends less on the reading step.

In our internal timing on test material, Lite runs took about 23 to 28 seconds. That is a measurement on our cases, not a guarantee. Timing depends on the length of your material.

Compare with Noetis Use it in the API

Three specialists

The seven Noetis roles are folded into three, so the slowest parallel read has fewer peers to wait on.

Terms and gaps reader

Obligations, durations, renewal and termination language as written, statements that cannot both be true, and documents referenced but missing.

Cost analyst

Figures, totals, escalators and thresholds. Arithmetic is rebuilt from sourced numbers instead of trusting a stated total.

Risk and questions reviewer

Liability, penalties and one sided remedies, cause and effect claims with their assumptions, and the questions worth asking before deciding.

What stays the same

Every quoted passage is located in your material with character offsets.
A separate review judges whether each passage supports the whole finding.
The brief is checked against reviewed findings, with one bounded repair pass.
Runs that do not pass the release check come back incomplete with reasons, and are not charged.

What it gives up

Three specialists instead of seven, so each one covers a wider brief.
Material up to 40,000 characters instead of 150,000.
The evidence use audit is off by default to save time in series.
Faster role models with a smaller thinking budget.

Choosing Lite

Add one field to your request. Everything else, including the response shape, polling and webhooks, is unchanged.

POST /api/public/v1/panel
{
  "question": "Should we renew this contract?",
  "material": "...",
  "engine": "lite"
}