GitHub ·
szczyglis-dev/py-gpt
Seeing what a Jev integration looks like as one line in a mature tool.
What it does
A desktop AI assistant covering a dozen providers and a long plugin list. Its Jev / System One (inline) plugin sends structured state to TypeSafe's model and gets typed decisions back — classification, routing, selection, verification, scoring — inside a chat session, documented beside image generation and voice as one more inline capability.
The repository's own description: “Desktop AI Assistant powered by GPT-6, GPT-5, Gemini, Claude, Grok, Ollama, DeepSeek, Perplexity, and more - chat, agents, tools, MCP, plugins, RAG, vision, voice, image and video generation, speech synthesis and speech recognition, web search, memory, presets, canvas, assistants, themes, computer use. Windows, Linux, and Mac.”
How it works
Plugins register extra tools inside the app's conversation loop, and the Jev plugin is one of them: an inline request can be answered with a label or a score instead of prose. Nothing about the model changes; what changes is that the app now has somewhere to put a decision.
The repository, by the numbers
Read from the GitHub API on 2026-09-27. Stars and the last push move daily — quote them with the date, the way we do.
What we checked
- the plugin list and the Jev / System One (inline) entry
- the plugin documentation page the README points at
- the licence field GitHub reports and the README's build stamp
- that the repository's own topic list includes jev
What we did not check. We did not install the application or exercise the plugin, so what it does with a real state object is what the repository documents, not something we measured.