Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up
Email triage

Sort an inbox with three typed questions: reply needed, topic, urgency.

A cited run processed 1,500 personal emails at ~200 ms each: a low “needs reply” probability auto-files, everything else reaches the LLM with its topic and urgency already decided. The LLM writes replies; it does not sort.

Get early access See the benchmark

In development — the benchmark numbers on this page are measured; the Decide-1 model is in training.

The problem

The LLM reads everything

Every newsletter costs the same tokens as the one email that matters.

Rules rot

Sender-based filters miss the important email from an unknown address.

Urgency is a scale

can-wait / this-week / today is a score, not a folder.

How it works with Decide
  1. 1

    Three questions per message

    needs_reply (noul), topic (choice), urgency (score 0–2).

  2. 2

    File the low end

    p(needs_reply) < 0.2 is filed without a model in the loop.

  3. 3

    Hand off with context

    The rest goes to the LLM with topic and urgency attached.

r = client.system_one(state=email, questions={
  "needs_reply": {"type": "noul",   "instructions": "This message needs a reply from me."},
  "topic":       {"type": "choice", "instructions": "What is it about?", "criteria": TOPICS},
  "urgency":     {"type": "score",  "instructions": "How soon?", "criteria": ["can wait","this week","today"]},
})
  • Noul — a calibrated yes/no probability
  • Choice — pick one option, a probability on each
  • Score — a position on your ordered rubric
Something you can check

Cited by atomicbot.ai and thepromptindex.com from a LangChain write-up: 1,500 emails at ~200 ms each on hosted Jev.

Numbers about Jev are its authors' or the cited source's; numbers from the benchmark are ours, on one harness, with the raw JSON published. Full table.

Get early access — and tell us which decision you'd hand over first

We're building Metatext Decide with the first teams that use it. Leave your email, answer as much as you like, and say yes to a short interview if you want a say in what ships.

  • Early access to Decide as it opens up
  • A 20-minute interview if you want one — research, not a pitch
  • The benchmark harness and raw results, today
  • No spam — early-access updates only
Where do decisions happen in your stack today?
What kind of decisions?
Have you tried Jev?
What stops you using a hosted decision API?
Would you pay?
Open to a 20-minute interview?

Research, not a pitch — we want to hear how you decide today.

Early-access updates only. No spam, unsubscribe any time.

Questions

Does it read my mail?
Self-hosted, nothing leaves your machine. That is the reason to prefer an open model for this one.

Related