flashrt-project/FlashRT

FlashRT

✅ Recently maintained

FlashRT is a high-performance realtime inference engine for small-batch, latency-sensitive AI workloads. The flagship integration is production VLA control for Pi0, Pi0.5, GROOT N1.6, and Pi0-FAST. Also support llm e.g, qwen3.6-27B

781Stars
103Forks
+257-day stars

C++ · Apache-2.0 · Last push 2026-10-28

Official repository & installation ↗
cudacuda-kernelsrealtime-inferencerealtime-vlagr00tgr00t-n1-6-3bpipi05vlaqwen

Star history

2026-04-272026-11-02
View historical data
DateStars
2026-10-04612
2026-10-05613
2026-10-06614
2026-10-07615
2026-10-08619
2026-10-09623
2026-10-10624
2026-10-11626
2026-10-12645
2026-10-13652
2026-10-14658
2026-10-15664
2026-10-16668
2026-10-17670
2026-10-18702
2026-10-19711
2026-10-20718
2026-10-21748
2026-10-22751
2026-10-23754
2026-10-24754
2026-10-25755
2026-10-26759
2026-10-27765
2026-10-28767
2026-10-29772
2026-10-30775
2026-10-31777
2026-11-01781
2026-11-02781

Get the project

git clone https://github.com/flashrt-project/FlashRT.git

Related projects & alternative candidates

Related domains do not imply feature parity. Review official documentation and licenses.

ollama

Run DeepSeek, Qwen, Gemma and other LLMs locally with one command, plus model management and an API.

★ 185,625orca

A workbench for running and orchestrating fleets of coding agents across desktop, mobile, and remote runtimes.

★ 112,825VoiceStudio

A fully local open-source voice studio for cloning, dubbing, transcription and audiobook creation across 646 languages.

★ 98,205