Skip to content

Case wall

Jev AI Infrastructure cases

29 public posts filed under Infrastructure. English summaries on this page, with a link to each case.

Last updated

These are the public posts filed under the Infrastructure category.

Filter

24

You asked for it and we delivered! We just posted the full recording of Fighting Fraud with Jev, our November 4 fireside chat with TypeSafe AI in San Francisco.…

Infrastructure

Sardine @sardine

EN

@sardine says the full recording of Fighting Fraud with Jev is up: the November 4 fireside chat with TypeSafe AI in San Francisco. Fraud and risk leaders spent the evening on why models built for chat wobble on approve-or-decline calls. Erik Gafni, TypeSafe AI’s co-founder and CTO, pinned the problem on RLHF, which trains a model to give people answers they like. His team uses RLCD to train Jev for calibrated calls that stay put when you ask the same question again. Soups Ranjan, Sardine’s co-founder and CEO, showed how that consistency pays off in sanctions screening. They built an AI engine that follows each bank’s SOP and now clears 90% of manual reviews at some of the banks they work with. It has not missed a true match in six months. Yoko Li, a partner at a16z Infra, moderated and pressed Soups on how they catch fraudsters who probe the models. AI shopping agents give themselves away by clicking the dead center of every field, with no natural mouse movement. Soups pointed to tells in their own hiring pipeline: up to 75% of the AI and ML candidates applying to Sardine show fraud signals, down to deepfakes in live video interviews. Erik urged anyone running AI millions of times a day to break the problem into small decisions and give the model as little to do as possible. They made Jev available as a decision node in Sardine Flow so fraud teams can build that way. The recording is at https://www.sardine.ai/events/fighting-fraud-with-jev-on-demand and the recap is at https://www.sardine.ai/blog/fraud-decision-model-stack.

You asked for it and we delivered! We just posted the full recording of Fighting Fraud with Jev, our November 4 fireside chat with TypeSafe AI in San Francisco.…

Poster from alginte software (@alginte_sw)

Infrastructure

alginte software @alginte_sw

EN

@alginte_sw says about 480 new CVEs arrive a day, and one in five arrives without a severity. They ask #TypeSafe #Jev to judge those from the description, inside a #KafkaStreams topology.

About 480 new CVEs a day. One in five arrives without a severity. We ask #TypeSafe #Jev to judge those from the description, inside a #KafkaStreams topology.

A pretty important new primitive just landed in 🤗 Transformers: **decision models**. I think this could matter much more than adding support for any single mode…

Infrastructure

Lysandre @LysandreJik

EN

@LysandreJik says decision models landed as a shared primitive in Hugging Face Transformers: DecisionConfig, pipeline("decision") and model.decide(...). He calls @typesafeai's Jev the OG: typed decisions, calibrated probabilities, one forward pass.

A pretty important new primitive just landed in 🤗 Transformers: **decision models**. I think this could matter much more than adding support for any single mode…

We are turning AgentRun into a fully featured, open source, managed agents platform. All the good stuff we did for our own harness like: - self-improving agents…

Infrastructure

Miguel Ríos Berríos @MiguelriosEN

EN

@MiguelriosEN says they are turning AgentRun into a fully featured, open-source managed agents platform, and lists @typesafeai's Jev and @p0's web-grounded workflows among the harness features.

We are turning AgentRun into a fully featured, open source, managed agents platform. All the good stuff we did for our own harness like: - self-improving agents…

Quoted @MiguelriosEN

Jev is now a native connector on Voximplant. Your code can ask it whether the caller finished the thought or actually agreed to the booking, and get a probabili…

Infrastructure

Voximplant @voximplant

EN

@voximplant says Jev is now a native connector, so code can ask whether a caller finished a thought or actually agreed to a booking and get a probability back in a few hundred milliseconds.

Jev is now a native connector on Voximplant. Your code can ask it whether the caller finished the thought or actually agreed to the booking, and get a probabili…

Poster from Bobo Li (@Bobo_hashtag)

Infrastructure

Bobo Li @Bobo_hashtag

EN

@Bobo_hashtag introduces Flide, open-source inference infra that batches many Jev-style decision questions about one artifact in a single pass.

Flide is open-source inference infra built for Jev-style decision models. Your agent asks Jev 18 questions about one article. Most engines run 18 prompts. Flide…

Poster from Matt Mastracci (@mmastrac)

Infrastructure

Matt Mastracci @mmastrac

EN

mmastrac runs EmbeddingGemma2 as a live /v1/systemone endpoint: ~20ms text, ~100ms images, distillable and multimodal.

It's been a while since I showed any Jev stuff but I've been cooking on a way to use @GoogleDeepMind's EmbeddingGemma2 as an insanely fast real-time /v1/systemo…

Poster from AI/ML API (@aimlapi)

Infrastructure

AI/ML API @aimlapi

EN

AIMLAPI MCP router: connect 10k+ tools/models, Jev picks the best for the task; claims ~96% cheaper than Claude.

Introducing Jev Router for Agentic Tools ⚡ 1. Connect your agent to 10,000+ tools and models in one MCP 2. @TypesafeAI’s Jev picks the best for your task 3. Res…

Poster from SuSu_酥酥👅 (@NFT_Chen)

Infrastructure

SuSu_酥酥👅 @NFT_Chen

ZH

@NFT_Chen announces DeepEval JevEval: Jev as a custom eval judge for Yes/No, Score, Choice, and agent traces.

💥我靠!LLM评估终于不用再靠另一个“爱飘”的生成式裁判了! DeepEval正式推出 JevEval:用Jev当自定义评估裁判,而不是每次都让另一个LLM去“生成”评分。 对比传统LLM-as-a-Judge: 🔹传统:模型自己解读标准 + 推理 + 生成结果(易飘、贵、慢、JSON还容易炸) 🔹JevEval:决策…

Jev from @typesafeai is now available as an evaluation model in @getbluejay_ai. Live today for voice, chat, and email agents. Eval every conversation, not a sam…

Infrastructure

Bluejay @getbluejay_ai

EN

@getbluejay_ai now offers Jev from TypeSafe as an eval model for voice, chat, and email agents, on every conversation.

Jev from @typesafeai is now available as an evaluation model in @getbluejay_ai. Live today for voice, chat, and email agents. Eval every conversation, not a sam…

Poster from Pi Changelog (@PiChangelog)

Infrastructure

Pi Changelog @PiChangelog

EN

@PiChangelog: Pi v1.4.0 lets codemode models.classify() accept images for Jev and Clef on typesafe, cloudflare-workers-ai, and OpenRouter.

Pi v1.4.0 is out! Highlights: - codemode models.classify() now accepts images for Jev and Clef models on typesafe, cloudflare-workers-ai, and OpenRouter - Compa…

Genkit Go 1.14 is out! - Typed human-in-the-loop tools - Improved model recovery from tool errors - New usage budgets and context compression - Decision model s…

Infrastructure

Genkit @GenkitFramework

EN

@GenkitFramework ships Genkit Go 1.14 with decision-model support for fast classification and routing (preview), plus HITL tools and budgets.

Genkit Go 1.14 is out! - Typed human-in-the-loop tools - Improved model recovery from tool errors - New usage budgets and context compression - Decision model s…

Poster from Zhining Liu (@zhining_liu)

Infrastructure

Zhining Liu @zhining_liu

EN

@zhining_liu launches JevAny: OSS toolkit to train/deploy System-1 decision models; 27B release beats Jev 1.13.0 on two evals.

1/5 🚀 Train and deploy your own Jev model with JevAny! We built an open source toolkit covering data prep, training, evaluation and deployment, plus a playgroun…

Poster from El Bruno 🇨🇦 🎗️ (@elbruno)

Infrastructure

El Bruno 🇨🇦 🎗️ @elbruno

EN

@elbruno released ElBruno.AI.Decisions, one .NET API for Choose, Score & Assess using Microsoft Foundry, local Ollama models, and Jev. Open source at https://github.com/elbruno/ElBruno.AI.Decisions.

Less chat. More decisions. 🎯 ElBruno.AI.Decisions: one .NET API to Choose, Score & Assess with Microsoft Foundry, local Ollama models and Jev. Probabilities, no…

Poster from Mastra (@mastra)

Infrastructure

Mastra @mastra

EN

@mastra lists recent launches including Classifiers with Jev, plus Factory integrations, memory hooks, and VPC-isolated Postgres.

Last week, we launched: 1/ Classifiers with Jev 2/ Jira, GitLab and https://incident.io integrations for Mastra Factory 3/ Memory hooks 4/ VPC-isolated Postgres…

Poster from Vercel (@vercel)

Infrastructure

Vercel @vercel

EN

@vercel says Jev is in the AI SDK for Python, tested by detecting Python vs English as text is typed and by writing Python one decision at a time.

Jev is now in the AI SDK for Python. To test it, we ran two experiments: ▪︎ Detecting Python vs. English as text is typed ▪︎ Writing Python, one decision at a t…

Poster from OpenRouter (@OpenRouter)

Infrastructure

OpenRouter @OpenRouter

EN

OpenRouter launched typesafe/jev-router, a cache-aware model router powered by Jev that picks model and reasoning effort per request.

Introducing typesafe/jev-router: a cache-aware model router powered by Jev and @typesafeai The Jev Router picks the best model and reasoning effort for each req…

1/ Last week, Jev - @typesafeai's System One model, was launched on OpenRouter. The interest was massive. We asked our community to find the most innovative way…

Infrastructure

OpenRouter @OpenRouter

EN

@OpenRouter says Jev, TypeSafe’s System One model, launched on OpenRouter the previous week. They asked the community for innovative uses of fast, cheap decisions and had Jev pick five winners.

1/ Last week, Jev - @typesafeai's System One model, was launched on OpenRouter. The interest was massive. We asked our community to find the most innovative way…

Poster from MotherDuck (@motherduck)

Infrastructure

MotherDuck @motherduck

EN

"@motherduck ships prompt_jev() SQL powered by TypeSafe Jev: 100k-row text classification in 40s for $0.50 vs 32 min / $37 for an LLM."

Text classification in MotherDuck just got ~50x faster at ~1% of the cost. prompt_jev() is a SQL function powered by Jev, TypeSafe's new system one model. 100k…

Poster from いのま (@iNoma_main)

Infrastructure

いのま @iNoma_main

JA

@iNoma_main released kojev, a library for calling Jev from Kotlin with answers as user-defined enums instead of strings. Missing choice descriptions are compile errors, and it targets JVM, Android, and iOS. The package is io.github.itisnomatter:kojev:0.1.0.

Jev を Kotlin から型安全に扱うライブラリ「kojev」を公開しました! ・答えが String ではなく自分で定義した enum で返る ・選択肢の説明は enum 定数に持たせるので、書き忘れはコンパイルエラー JVM / Android / iOS 対応 io.github.itisnomatter:k…

Poster from Vercel Developers (@vercel_dev)

Infrastructure

Vercel Developers @vercel_dev

EN

@vercel_dev says Jev by TypeSafe is free on Vercel AI Gateway until Sept 25, and links the Gateway model page.

Jev by @typesafeai is free on Vercel AI Gateway until Sept 25. Build with the fastest adopted model on the Gateway at no cost. https://vercel.com/ai-gateway/mod…

Quoted @vercel

Poster from Vercel (@vercel)

Infrastructure

Vercel @vercel

EN

@vercel says Jev was adopted faster than any other model in AI Gateway history. On the first day TypeSafe reached about 13% of teams, 2x the GPT-5.6 family and 6x Fable 5.1.

Jev was adopted faster than any other model in AI Gateway history. In the first day, @typesafeai reached ~13% of teams, 2x the GPT-5.6 family and 6x Fable 5.1.

Poster from OpenRouter (@OpenRouter)

Infrastructure

OpenRouter @OpenRouter

EN

@OpenRouter says Jev by TypeSafe is on OpenRouter in beta. It takes app state plus a typed question and returns a typed decision with a probability, with no JSON prompting or parsing layer.

Jev by @typesafeai is now on OpenRouter, in beta. Jev is a System One model. Instead of generating text, it takes your app's state plus a typed question and ret…

Quoted @thread

Unofficial developer resource. Not affiliated with TypeSafe AI. These posts belong to their authors. jevnotes.com does not host the videos.