Featherless.ai opens up Jev with launch of industry-first converter to bypass closed AI systems and ‘read’ images

San Francisco, USA; 29th September 2026: Featherless.ai has released Simple Jev, an open-source library designed to convert open-source AI models into high-speed, zero-shot classification engines. 

While Jev was previously a closed-source system, Featherless.ai’s Simple Jev extends structured decision functionality to open-source models, enabling applications to evaluate incoming data and return categorical assignments or binary choices without generating conversational text. 

Simple Jev also includes industry-first vision support for decision workflows. Currently only available with Gemma or Qwen models, this feature enables developer teams to use images as context for their queries for instant classification. By bypassing conversational text the solution minimises latency and cuts down on compute usage. 

Featherless.ai is introducing free public endpoints which will enable developers to use them for testing classification models without the need for API keys and logins. For production tasks that require longer context windows and smaller models, the inference pricing starts at $0.03 per million input tokens, while output tokens are free.  

Simple Jev is available now as a fully open-source library through GitHub, providing developers a dedicated model distillation workflow to compress complex prompts and decisions from large frontier models down to smaller, highly efficient open models for production. 

Eugene Cheah, CEO and co-founder of Featherless.ai, said: “For too long the AI industry has focused almost exclusively on complex reasoning by creating big models for slow, multi-step analysis. However, in software engineering, you will often have tasks that can benefit from making quick decisions. Running an application through chat models just to route tickets and categorise files is a waste of resources and computing power. It works well for token maxxing, but we’re focusing more on token minisation. And this is where Simple Jev comes in, as it gives developers control over fast decision-making processes with open-source models, substantial speed, cost advantages and the option to have vision support. 

“The broader market is primarily focusing on monolithic generalist models, failing to see the enormous value in specialised and customised open-source models for particular tasks. The future and its efficiency will be driven by the millions of lightweight, dedicated models, which perform significantly faster and much cheaper than any closed solution. Our aim with Simple Jev is to help dev teams and businesses optimise their unit economics without vendor lock-in.”