Jev for beginners: how to use it and what to build

September 28, 2026

AI Summary

5 min read

The host of How I AI opens not with a general overview but with a specific claim: Jev, a new model from TypeSafe AI, has become the most explosively useful tool in his stack, not despite being limited, but because of it. Unlike frontier models that generate long strings of text, Jev takes text in and returns only "type safe values out"—predefined choices, scores, or yes/no probabilities. It costs four cents per million input tokens and charges nothing for output because it barely outputs anything. The host has spent less than ten dollars on Jev tokens while building features he says would have cost hundreds of thousands of dollars three years ago.

What Jev actually does

Jev returns exactly three kinds of values. A choice picks one option from a list you supply—dress, jeans, or ski gear for a date outfit. A score ranks something on a scale you define, like triaging a bug as cosmetic, broken, or blocking. A boolean (called "null") returns the likelihood that the answer to a question is yes—99% chance that the host is a podcaster. These are not sophisticated outputs, but the host argues that 90% of software engineering is exactly this: making a decision, routing something, scoring something, saying yes or no. Jev is, in his framing, a "high agency decision model" that replaces smart if-statements in code. It is fast enough to run in real-time loops where other L

Continue reading the full summary in the app — free to try.

Read Full Summary →

Free • No credit card required

Listen to Audio Summary Open in App

Never miss an episode of How I AI

Get every new episode summarized in your inbox — free, ~5 minutes to read.

No spam. Unsubscribe anytime.

What you'll learn

1 (00:00) **Introduction: Why Jev is the Model You Should Care About** - The host sets up Jev as the standout model from a week of new AI releases, calling it a fast, cheap, system-one decision model from TypeSafe AI that has exploded personal productivity and product use cases.
2 (02:49) **What is Jev? Text in, Type-Safe Values Out** - The host explains the core mechanic: unlike standard LLMs that output strings of generated text, Jev takes text in and returns pre-defined, type-safe values like choices, scores, or booleans.
3 (03:40) **Jev is Cheap and Fast AF** - A breakdown of the cost and performance advantages over standard LLMs.
4 (07:39) **Use Case #1: PR Analysis for CTOs and Product Leaders** - The host demonstrates a real application that would have cost $100,000 three years ago, now done for pennies.
5 (11:12) **Use Case #2: Analyze Your Own Local Code Sessions** - A use case anyone can run on their local machine, even without a big repo.
6 (12:55) **Use Case #3: Personal Gmail Triage** - The host briefly describes running Jev on their personal inbox.
7 (14:21) **The Big Picture: Building a Product Insights Graph with Jev** - The host reveals how Jev unlocked a feature that frontier models failed at, costing thousands of dollars.

+ Full timestamped outline available in the app

Show Notes

Jev is TypeSafe AI’s new decision model. It returns type-safe structured values (a choice, a score, a probability) instead of generated text, at 4 cents per million input tokens with no output charge. This week I ran it on five real projects: PR categorization, a meta-analysis of my own Claude and Codex sessions, Gmail triage, the ChatPRD product insights graph, and a live audience dashboard built from 4,500 YouTube comments.

What you’ll learn:

What makes Jev fundamentally different from every other model I’ve used
How I analyzed 1,700 PRs for 9 cents and what I found out about where my engineering effort actually went
The personal meta-analysis you can run on your own Claude and Codex sessions right now
Why I stopped using Jev alone, and what I pair it with now
How I turned 4,500 YouTube comments into a searchable audience dashboard for almost nothing
The real-time app I built in an afternoon that shows something surprising about Jev’s speed
Why Jev’s pricing model is different from any LLM I’ve used, and what it makes practical to build
The ChatPRD product insights project: 1,100 signals, 200,000 classifications, and what it cost me

—

Brought to you by:

OpenArt—An all-in-one AI creation platform for images, videos, music, audio, and more

—

In this episode, we cover:

(00:00) Jev launch and what makes it different from every other model

(02:49) Type-safe values explained

(05:28) Understanding Jev outputs

(07:39) Use case 1: PR categorization and pairwise clustering

(11:12) Use case 2: analyzing your own local Claude Code and Codex sessions

(13:00) Use case 3: Gmail triage with Jev scoring and LLM follow-up

(14:30) Use case 4: ChatPRD’s product insights graph