Jev in Practice: Browser Use and Agent Workflows

A new tool called jsort ranks arbitrary lists from a natural-language description, for example ordering 95 Fed statements by how hawkish they are on inflation, with this week's Warsh statement at 15th.

A roundup of five projects and resources built with Jev, covering browser agents, search tools, driving simulations, and experiments that bring similar decision-making to open models.

The author says using Jev to speed up Search Console and PostHog reads, citation analysis and related steps cut per-client SEO/GEO fixes from about $250 to a tenth of that.

Jev takes state and typed questions and returns branchable decisions with probabilities through three primitives, Choice, Score and Noul, and is listed on Vercel AI Gateway.

An Arabic post explains the Jev pattern: the LLM generates, Jev judges and classifies, and code executes, moving frequent small decisions out of the large model.

The author built a single-file HTML tool with Codex where pasted text is classified sentence by sentence by Jev as claim, evidence, guess, or instruction.

A reminder that Jev (typesafe-ai/jev) is free on Vercel AI Gateway until 9.25, with no cost for input or output, for agent routing, ticket triage, and guardrails.

Aile routes every request to a lender, comparing price first and using health to break ties, as a typed decision with a probability, and is now wiring it into its harness.

A developer built a bridge-playing bot called Jev Bridges on TypeSafe's new Jev model and asks whether anyone has tried the same.

Ryze AI uses Jev to score competitor pages, decide which page elements to keep or change, and rate how likely a page is to be cited by AI, available in its app and via MCP/Claude.

The author uses Jev to score a task in about 100ms, pick one of N models and return a confidence score plus a needsReview flag, escalating only uncertain requests to Claude or GPT.

Jev returns probability scores for decisions like fraud risk, review or approval in 70-500ms, aimed at high-volume small decisions such as risk checks, recommendations and agents.