DOQS

An API that scores document quality and structure, available in tiers.

DOQS scores document quality and structure through an API, with tiered access.

Stop guessing which layers to prune.

The problem with pruning

How DOQS works

What the API returns

Simple, honest pricing

Free

Regular

Pro

DOQS takes your model profile and budget constraints and returns a deterministic pruning schedule, with predicted cost savings, a signed audit certificate, and calibrated error bounds. One API call. No experiments.

Deterministic · Auditable · Live API at doqs.onrender.com

Optimize calls are the only thing that counts against your limit. Profile submissions, evaluations, audits, and schedule retrievals are always free.

All plans include access to /v1/evaluate, /v1/audit, and /v1/schedules at no additional charge. Prices in GBP. Cancel any time.

Free tier available immediately. No credit card required.

• Per-layer retention schedule
• Exact fraction of each layer to keep. Not a uniform ratio, each layer gets its own number based on its position and sensitivity.
• Predicted cost savings
• How much cheaper each 1,000 inferences will be after pruning, in your declared currency unit.
• Economic breakeven point
• How many inference blocks until the optimization cost pays for itself. In most cases: 1 block.
• Surrogate corrections
• The API tells you explicitly how accurate its own predictions are, with a corrected latency figure for deployment planning.
• Signed audit certificate
• A verifiable STRICT-PASS certificate stored with the schedule. Prove to anyone that constraints were satisfied.