Decision models: the cheap, fast step inside an agent
Most steps in the recipes here are not writing. They are choices: is this email a refund request, is this lead a fit, can this order ship. A decision model answers those closed questions in milliseconds for a fraction of a cent, and leaves the writing to a language model.
What it is
Jev is a hosted model that takes some text plus a typed question (yes or no, pick one of these, score from 1 to 5) and returns an answer with a confidence score. It never produces prose, which is why it is fast and cheap. The community directory Made with Jev catalogues more than 800 public builds, and most of them use it for the same few jobs.
The pattern those builds share fits the way every recipe here is built: a language model writes, the decision model decides, code acts, and a person signs off on anything customer-facing, financial or hard to undo.
Where it fits in these recipes
| The job | A published build | Author's figure | Recipe here |
|---|---|---|---|
| Triage an inbox into buckets | 500 emails classified (Riley Brown) | 3.5 cents total | Support ticket triage |
| Score leads and outreach before sending | 700 leads scored (Romàn) | 40 s, $0.09 | Inbound lead qualification |
| Screen for fraud, escalate the unsure ones | 100 emails, low-confidence cases sent to a larger model (Hassan) | 96/100 correct, about $0.07 | Order-exception triage |
| Classify a stack of documents | 26-sheet plan set (Trinay Hari) | 2.9 s, $0.0052 | Receipt categorization |
| Break down competitor ads | 724 ads from 37 brands (Matthew Berman) | about 40 s, $0.09 | Content repurposing |
| Pick internal links for SEO | Internal linking tool (Ian Nuttall) | Up to 500 pages | No recipe yet |
| Ask a question of every row in a database | jev() for PostgreSQL (Zachi) | 129 rows in about 1 s | Weekly KPI digest |
| Check an agent's action before it runs | Agent safety monitor (Nick Ma) | Caught most test attacks, few false blocks | Any recipe with a send or pay step |
Every figure is the one its author published on the linked page. We have not re-run them.
New: decisions about images
Jev 1.14, released on 14 October 2026, added image input (PNG or JPEG) and a "Box" answer that points at one region of an image, such as the signature line on a scanned form. Images are billed as ordinary input tokens. TypeSafe AI labels both features as preview, not production-ready, so test them on your own sample images first.
That opens up checks that used to need a person or a large model to look at every photo: sorting receipt photos, checking product photos against marketplace image rules, flagging damage in return photos, confirming proof-of-delivery photos, and checking that a form is signed.
Is your step a good fit?
The answer is a fixed set of options
If you can list the possible answers before the call (refund, exchange, question, spam), it fits. If the step needs a written reply, keep it on a language model.
The same decision repeats
The savings are per item. Hundreds of emails, orders or leads a week add up; one decision a day does not.
A wrong answer can be caught
Act on the confident answers and send the rest to a person or a larger model. Start with actions you can reverse.
Caveats
- Jev is a paid, hosted API from TypeSafe AI. Check its current pricing and terms before you build on it.
- Image input and the Box answer were labelled as preview features when this page was written (November 2026).
- Keep the human sign-off points in each recipe. A fast model makes a bad decision faster too.
- Agents Lookup is not affiliated with TypeSafe AI or Made with Jev.
Want a decision step added to a recipe?
Setup help covers wiring a triage or scoring step into any recipe in the directory, with the confidence threshold and fallback set up for you.