Jev: Typed Probabilistic Decisions
Jev is a model that provides typed probabilistic decisions instead of generating text, designed for routing, scoring, and real-time agents.
Use cases
Jev use cases and projects
Explore how developers use TypeSafe AI's Jev for routing, classification, agents, search and more — scored builds from the live radar.
01 · Routing
Routing
135 builds
Routing projects use Jev to send work to the right model, tool, or path — fast typed decisions instead of free-form generation.
View routing projects
Jev: Typed AI Agent Decisions
Jev is an AI model that takes shared state and typed questions, providing direct code-usable answers with choices, scores, and probabilities, enabling efficient decision-making for AI agents.
JEV x Automation Room
This project explores giving an AI model, JEV, a programmable robotic environment to reason about tasks, allowing it to observe the consequences of different decisions in a simulated workspace.
Jev: AI Agent Decision Orchestration
Jev acts as a decision-making layer for AI agents, choosing the next move by splitting intelligence from execution and enabling parallel processing of decisions.
02 · Classification
Classification
504 builds
Classification builds lean on Jev Choice, Score and gate-style judgments so apps can accept, reject, label or pick options with typed confidence.
View classification projects
Vectorless RAG with Jev
This project introduces a novel approach to Retrieval Augmented Generation (RAG) that replaces traditional embedding methods with a structured, table-of-contents-based navigation system, significantly improving speed and reducing costs by offloading non-generative tasks to Jev.
PostgreSQL Extension for TypeSafe AI Jev Integration
This project is a PostgreSQL extension that allows SQL queries to directly call TypeSafe AI's Jev for classifying rows within a database table, avoiding the need to export data for external processing.
03 · Agents
Agents
324 builds
Agent projects use Jev inside loops and toolchains: choosing the next action, gating tools, or controlling what runs when the environment changes.
View agents projects
Jev Decision Layer for Voice Pipelines
Jev acts as a decision layer in voice pipelines, utilizing typed primitives like Noul (yes/no with probability), Choice (labeled decision), and Score (scalar on a defined rubric) to process conversational turns before and during LLM responses.
Jev is what happens when you stop forcing AI to talk.
Jev is what happens when you stop forcing AI to talk. It’s the first public “System One” model from TypeSafe AI, built for fast decisions inside software—not conversations. You give it: • Unstructured state: text,
Jev: Typed Probabilistic Decisions
Jev is a model that provides typed probabilistic decisions instead of generating text, designed for routing, scoring, and real-time agents.
04 · Search & Retrieval
Search & Retrieval
42 builds
Search and retrieval builds use Jev to match intent, rank candidates, or decide what context belongs in the next step.
View search & retrieval projectsVectorless RAG with Jev
This project introduces a novel approach to Retrieval Augmented Generation (RAG) that replaces traditional embedding methods with a structured, table-of-contents-based navigation system, significantly improving speed and reducing costs by offloading non-generative tasks to Jev.
AI News Feed Rebuild
This project rebuilt an AI knowledge graph into a news feed using a decision model for classification, resulting in a significantly cheaper and faster system.
TypeSafe AI Use Case for Mac App Troubleshooting
This project integrates TypeSafe AI into a Mac app to provide setup and troubleshooting assistance by reading user questions and referencing the built-in manual to identify relevant documentation or status.
I asked an AI 52,350 questions.
You give it a page and a list of questions, and it sends back a yes, a no or a score for each one, with a probability. 25 judgments about a page take about a third of a second.
05 · Developer Tools
Developer Tools
90 builds
Developer-tool projects put Jev behind editors, CLIs, evals and infra so typed judgments show up where engineers already work.
View developer tools projects
Jev: A Judgment Layer for Development
Jev acts as a judgment layer in the development loop, sitting between the developer's assessment of completion and handover to a maintainer, identifying potential issues and inconsistencies.
Jev: Rule Selection for Claude Code
Jev intelligently selects applicable rules for Claude Code based on prompts and file paths, optimizing context token usage and ensuring relevant rule delivery.
06 · Automation
Automation
146 builds
Automation builds keep Jev in the loop — controlling agents, gating steps, and choosing what happens next as state changes.
View automation projectsJev Decision Layer for Voice Pipelines
Jev acts as a decision layer in voice pipelines, utilizing typed primitives like Noul (yes/no with probability), Choice (labeled decision), and Score (scalar on a defined rubric) to process conversational turns before and during LLM responses.
TypeSafe Jev Desktop Interaction
This project enables computer use within a Dockerized Linux environment by using the accessibility tree as 'eyes' and TypeSafe Jev to answer typed questions for the next action, with a11y invoke or xdotool serving as 'hands'.
Jev: A Decision-Making AI Model
Jev is an AI model that provides typed decisions, probabilities, and confidence scores for software to act upon, differing from traditional LLMs by not generating answers token by token.
