Pycharm implementation

[Karbonite]

Has anyone successfully implemented this on Pycharm? I have a Pro license , added the Codex to allow Open-AI Compatible connections, pointed the base-url to my local, test said 'connected successful', the only DeepSeek models fully accepted but when trying to test it just the chat just spins until timeout. Do env varibles need to set, or am I missing something else?

Using Windows via Desktop Install from website.

[tashfeenahmed]

JetBrains AI Assistant does not talk to the OpenAI compatible endpoint the way Codex does, which is why the connection test passes but the chat never returns. It needs the Ollama emulation surface instead, and that is off by default.

Turn it on under Keys → Agents. On a normal desktop install pick open-loopback. Then point JetBrains at http://localhost:3001 as an Ollama server rather than as an OpenAI provider.

One thing to watch on Windows.... if you are running the server in Docker rather than the desktop app, open loopback will refuse the connection because the socket peer is the Docker bridge IP and not loopback. In that case use key-required and send your unified key as a bearer token.

Details are in docs/clients.md (https://github.com/tashfeenahmed/freellmapi/blob/main/docs/clients.md#ollama-clients). Let me know if it still spins after that and I will dig in.