whisper_ros

A set of ROS 2 packages that bring whisper.cpp speech-to-text into ROS 2, using audio_common for audio and Silero VAD for voice activity detection.

Licenses

Code: MIT (opens in a new tab)
Read by Robotheca from LICENSE (opens in a new tab) (MIT) at commit 4e0f285 (opens in a new tab), on 16 Oct 2026.
Commercial use: Yes

Runs Whisper for speech-to-text transcription inside ROS 2, with a listen action for sending transcription goals.
Uses Silero VAD to detect voice activity in the incoming audio.
Offers many parameters covering the model, sampling, transcription, token timestamps, decoding fallback, diarization, GPU and DTW timestamps.
Can be built with CUDA for GPU inference on Whisper and Silero, or with OpenBLAS for faster CPU-only inference.
Includes a Docker image, with optional CUDA support, and launch files plus a demo node for trying it out.
Supports ROS 2 Humble, Iron, Jazzy, Kilted, Lyrical and Rolling.

Written by Robotheca from the project's README (opens in a new tab) at commit 4e0f285.

Difficulty: Intermediate
Hardware it needs, as the project states it: For CUDA: the CUDA Toolkit, plus cuDNN to run Silero VAD with ONNX and CUDA; the NVIDIA Container Toolkit for CUDA in Docker. CPU-only inference can be sped up with OpenBLAS. No other requirements are stated.
How to start: The project's website (opens in a new tab)

Checked on GitHub on 16 Oct 2026.

No robot on Robotheca is linked to it yet. Links come from a robot's robot.yaml, from its dependency files read as text, or from Robotheca's editors.

Robotheca lists software as a reference: nothing is hosted or run here, and every fact links to where it was read.