best use of local LLMs
all competitionsFor the project that ran its AI on-device or self-hosted, not just an OpenAI API key in a .env file.
₹2,000 worth of prizes.
Submission deadline: September 21, 2026
submit
Submissions are closed.
The deadline for this competition has passed - we're no longer accepting entries.
what we mean
A model running locally or self-hosted (Ollama, llama.cpp, a local inference server, etc.) doing real work in your project - not just calling a hosted API. This also covers lightweight AI models or computer vision running directly on microcontrollers, low-power microchips, or single-board computers (ESP32, Raspberry Pi) - smart intelligence without the cloud.
how to participate
Submit your project through the Hub app as usual, then come back here, search for it below, and enter it for this prize. Every submission using a local/self-hosted LLM is eligible - entering here is what puts it in the running for judging.
- —The LLM actually runs locally/self-hosted, not through a third-party hosted API.
- —Submitted through the Hub app like every other project, then entered for this prize below.
- —Mention the model and how it's run in your README, so it's easy to verify.