# microsoft/onnxruntime-genai

> A C++ library for running generative AI models (LLMs) on-device via ONNX Runtime: it implements the generative loop, pre/post-processing, KV cache management, search and sampling, and grammar constraints for tool calling. APIs are available for Python, C#, C/C++, and Java.

- Magnitude: 0.0 out of 10 — Quiet
- Stars: 1,127 total · +0 stars on 2026-09-28 (UTC), ≈ 0 by evening
- Star trust: star growth looks organic
- Category: Language models · Language: C++ · License: MIT · Created: 2023-11-13 · Last push: 2026-09-23
- GitHub: https://github.com/microsoft/onnxruntime-genai · Page: https://gitnova.dev/en/r/microsoft/onnxruntime-genai

## Useful for

- Run local LLMs like Phi, Llama, or Qwen on CPU without cloud services
- Build a desktop or mobile app with offline text generation in C#, Python, or Java
- Constrain model output with a grammar to enable tool calling

## Why it’s here

- A steady background without notable changes.

## Star trust

Star growth looks organic.

Star-trust labels are heuristics based on the repository’s behavior, not a check of every stargazer.

## Numbers

- Forks: 356
- Issues and pull requests: 2,481
- Watchers: 48
- Average over the last week: 0 per day
- Usual pace: 1 per day
- Latest release: v0.16.0 (2026-09-22)

## Stars per day, last 30 days (oldest → newest, today is partial)

2026-10-04 … 2026-11-02: None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None, None

## Magnitude by day, last 12 days

09-17 0.4, 09-18 0.8, 09-19 0.9, 09-20 0.6, 09-21 0.6, 09-22 0.1, 09-23 0.1, 09-24 0.0, 09-25 0.0, 09-26 0.0, 09-27 0.0, 09-28 0.0

---

Magnitude (0–10) measures how fast and how unusually interest in a repository is growing right now. It is not a quality score.
Days are UTC. “So far today” is a fact; “expected by the end of the day” is a forecast.
Summaries and use cases are written by an LLM (DeepSeek V4.1 Flash) from the README and may be inaccurate: verify specific claims (benchmarks, speed, hardware) in the repository itself.
Data as of 2026-09-28 10:29 UTC, updated every 30 minutes.