Sediment 1

Institutional memory for your company.

A model trained on your company’s own documents. Ask it what you decided and why, get smart on a new set of documents in an afternoon, and draft in your own voice.

What it does

  • Research. Load a set of documents and ask questions across them and your notes instead of reading everything twice.
  • Memory. Find everything the company has looked at before, what it decided and why, and what something new resembles.
  • Deliverables. Draft first-pass memos from your own prior work, in your own voice.

Accuracy by task family

  • Sediment 1
  • Opus 5
  • GPT-5.6 Sol
025507510080.070.070.0Single-fact recall90.093.363.3Absence and abstention83.654.453.3Convention conformance

Speed and cost

AttributeSediment 1Opus 5GPT-5.6 SolClaude Code
Mean cost per query$0.013$0.25$0.09$0.48
Median latency per query20.1s40.5s51.8s24.8s

Your data

Each company runs on its own dedicated, isolated instance. Your documents are never pooled with another company’s and never used for any other customer.

Who it’s for

Built for teams of 5 to 30, no IT department required.

Book a demo