Case wall
Jev AI Agents cases
79 public posts filed under Agents. English summaries on this page, with a link to each case.
Last updated
These are the public posts filed under the Agents category.
No posts match these filters.
@KaranVaidya6 introduces Composio Firewall. Every agent action is evaluated with Jev and custom SLMs; it checks actions and information coming back. Live at composio.dev/firewall.
Agents have a trust problem. Today’s models can do most work. What they need is the right boundaries. In traditional software, those were set at build time. Wit…
@allietheicon is building responsive voice interactions with a custom STT and TTS pipeline. Jev routes intents to narrow handler paths that respond quickly.
I’m on the custom personal agent train for sure, but my particular obsession is extremely responsive and expressive voice interactions. Fully custom STT and TTS…
@RealMarvelX shows Beacon + Jev: one coding agent’s lessons become shared memory for other agents.
your AI coding agents can now learn from each other’s mistakes instead of repeating them. Beacon + Jev lets you turn useful lessons from one agent’s coding sess…
@MiguelriosEN open-sources Pi-vision: TypeSafe Jev analyzes images/video/docs while a Pi agent orchestrates, cutting frontier-token spend.
A few weeks ago, we built a new harness, AgentRun, which turns repetitive steps into workflows and code using Jev. This week we're open-sourcing its best featur…
@Xavidop built a Kubernetes software factory: Gemini and Genkit build each stage, and Jev gates ship, rework, or ask a human.
I built a software factory on Kubernetes 🏭 Feature request in, tested and reviewed Go module out. Gemini (@googleaidevs) builds each stage as a @GenkitFramework…
@abgregs built a Claude Code skill router plugin: each skill scored with a Jev Noul before the model sees the prompt.
My experiment building a skill router plugin for Claude Code with Jev: https://abgregs.github.io/jev-skill-router/ How it Works Before the model sees your promp…
Quoted @CompleteSkeptic
@Su1eyman11 tries reasoning plus inverse kinematics with no learned motion policy: Jev as System 1, Astra as fallback, IK for motion, fruit into a tray.
How far can we get with reasoning + inverse kinematics, without a learned motion policy? I built this small experiment with Jev as System 1 and Astra as the fal…
@localixai explains decision models (Jev from TypeSafe) and shows where they built them into a pet project.
Decision models (Jev from TypeSafe) are a relatively new class of models: they don't write text, they make decisions. I built them into my pet project in a few…
@PeteCapeCod ships npm dev-orchestrator: Jev/OAI Decisions/CF Clef drive Codex CLI or Claude Code instead of expensive reasoning tokens.
I've been working on a new thing! It's a cli that uses other clis, basically you can use Jev (or OAI decisions, or CF's Clef/flash) to make calls either codex c…
@parkerbrowndev hooks Jev before tool calls in a Pi/Harbor setup with Haiku; reports better safety, half the tokens, higher API $ from broken caching.
Jev pre-tool call router experiment pass 1 is in! Design - Haiku (fast) model - Pi harness - Harbor eval - Jev gets hooked in BEFORE tool calls Result: - MUCH b…
@sinkush: a local hybrid pass takes the top 30, then Jev judges each hit in parallel in about 0.3 s for a tenth of a cent.
Semantic search isn't relevance search. I first built agentic, file-system-style retrieval: accurate, but ~30 s per question. Now a local hybrid pass finds the…
@typakon4 shipped jev-layer 0.3.0 with Jev 1.14 screenshot input, visual grounding, and opt-in desktop action advice for agent harnesses.
Shipped jev-layer 0.3.0 with Jev 1.14: screenshot input, visual grounding and opt-in desktop action advice for agent harnesses. Jev recommends. Your host keeps…
@akshay_pachaar put Jev in TrueForge for client follow-up matching: as accurate as Opus on 40 requests at 150x lower decision cost.
i just put Jev inside an open agent harness. (150x cheaper decision-making than LLM) agents burn a lot of tokens on decisions that are really just picking one o…
@CaineArdayfio demos Thud: Mobile Use plus Jev audio that maps spoken commands to page actions without transcription.
Mobile Use + Jev for audio = ULTRAFAST⚡️ 10x faster and 20x cheaper than Gemini Live. It's our Audio Action model, Thud. Thud listens and then does. It has a pr…
@smokyproductco open-sourced agy-cua: background computer use for Google Antigravity via Cua Driver, built with @trycua and TypeSafe Jev.
Do you use @antigravity ? Do you wish antigravity had computer use? Have you tried the Google Computer Use feature gated behind API and realized it sucks ass? T…
@Sumanth_077 built a data-analyst agent where Qwen writes SQL/answers and Jev reviews question clarity, SQL fit, and grounded explanations; open-sourced under Hands-On-AI-Engineering.
I built a data analyst agent with Jev as the review layer! Qwen handles the main agent work, Jev acts as the review layer, and Vercel eve runs the agent loop, t…
Quoted @Sumanth_077
@k_cieslak built an OpenCode plugin that uses TypeSafe Jev for automatic handoff and routing between harness agents, enabling Plan → Build → Review without human interaction. Repo: https://github.com/Krzysztof-Cieslak/jev-agent-control
I've built a small tool - @opencode plugin using @typesafeai Jev to do automatic handoff and routing between agents available in the harness. It basically enabl…
@jetwaniavinash built jevmem so Claude Code remembers project rules: say it once, it writes JEVMEM.md, and Claude sees it next session.
I kept typing the same rules to Claude Code. "Dates are day-month-year." "Don't push without asking." So I built jevmem. Say it once, it writes it down in your…
@NickGideo published question-kit, a developer kit for building on Jev / System One models, and linked https://github.com/Engid/question-kit.
i'm bullish on typesafe and Jev! so I made question-kit, to give developers a kit for building on Jev. go check it out! https://github.com/Engid/question-kit
Quoted @CompleteSkeptic
@dbreunig sketched a model router in about 15 minutes with DSPy and Jev that picks from the initial prompt, and linked a gist.
A quick example of a model router, using DSPy and Jev, based on your initial prompt. Sketched this out in ~15 min, by request, and am surprised at how well it w…
@shawnbuilds made a custom Jev memory harness for Claude Code mods, using TypeSafe via treg, and says the rebuild prompt is below the post.
Claude Code Mods make Claude 100% customizable, so I built a custom Jev memory harness built using @typesafeai via @treg_ai thanks to @trq212 for the idea! full…
Quoted @trq212
@MadisonJRickert built a Jev classifier for Claude Code auto mode and says the permission checks it decides are 2x faster.
Anthropic released native mods for customizing Claude Code yesterday, so I did the obvious thing and built a Jev-based classifier for auto mode! 2x faster permi…
"@N01ennn shows a RAG loop where one Jev call scores 20 passages + answerable/injection gates, then a second Jev call grounds claims."
this is pure f*cking treasure A Stanford AI research group finally drew the perfect RAG system: retrieval, Jev and agents in one loop, and it fixes the 3 things…
@DominikTornow's stfu asks Jev on each prompt how long Claude's reply should be, then adds that limit to the turn.
Shut up Claude with @typesafeai. On each prompt, Jev decides how long the reply should be and adds a length limit to the turn. Get an answer, not an essay. http…