What it does
Core capabilities at a glance
- Real-time STT on CPU and GPU
- Supports large-v3, tiny through large Whisper models
- Core ML and Metal acceleration on Apple Silicon
- OpenVINO and CUDA support
- Speaker diarization via pyannote integration
- Streaming audio transcription with VAD
Deep dive
The full breakdown - performance, comparisons, and setup
Whisper.cpp
Whisper.cpp brings OpenAI's Whisper speech-to-text models to local hardware with optimized C/C++ inference that runs faster than real-time on modern CPUs.
What it is
Whisper.cpp is the C/C++ implementation of OpenAI's Whisper speech-to-text models, created by the same developer behind llama.cpp. It converts spoken language into text with accuracy rivaling cloud services.
Performance you'll see
| Hardware | Model | Speed (relative) | VRAM |
|---|---|---|---|
| Apple M4 Pro | large-v3 | 12x real-time | CPU only |
| RTX 4090 | large-v3 | 25x real-time | ~2GB |
| Intel i7-13700K | medium | 8x real-time | CPU only |
| Raspberry Pi 5 | tiny | 2x real-time | CPU only |
How it stacks up
| Whisper.cpp | Sherpa Onnx | OpenAI Whisper API | |
|---|---|---|---|
| Speed | Fastest | Fast | Cloud |
| Accuracy | High | Medium | Highest |
| CPU inference | ✓✓ | ✓ | Cloud |
| Language support | 100+ | 20+ | 100+ |
| Self-hosted | ✓ | ✓ | ✗ |
Get started
git clone https://github.com/ggerganov/whisper.cpp.git
cd whisper.cpp
make -j
# Download model
bash models/download-ggml-model.sh large-v3
# Transcribe
./main -m models/ggml-large-v3.bin -f audio.wavWhen to use something else
- Need STT + TTS + VAD in one SDK: Sherpa Onnx
- Multi-modal API server: LocalAI bundles Whisper
- Higher accuracy: Use OpenAI Whisper API or larger models
Frequently asked
Quick answers to common questions
What is Whisper.cpp?
Whisper.cpp is a tts-stt tool for local AI workloads. High-performance, real-time speech-to-text using OpenAI's Whisper models in pure C/C++.
Is Whisper.cpp free and open source?
Yes, Whisper.cpp has 50,535 GitHub stars and is licensed under MIT. You can self-host it for free on macos, linux, windows.
What platforms does Whisper.cpp support?
Whisper.cpp runs on macos, linux, windows.
What hardware do I need for Whisper.cpp?
The hardware requirements depend on which models you run. Check our hardware directory for compatible GPUs and systems. Whisper.cpp has 50,535 GitHub stars and an active community.
Does Whisper.cpp support GPU acceleration?
Whisper.cpp supports GPU acceleration via CUDA, Metal, or Vulkan depending on your platform. For the best performance, pair it with an NVIDIA RTX 4090 or 5090.
What are the best alternatives to Whisper.cpp?
Popular alternatives include other tts-stt tools in our directory. Browse our full collection at /tool for comparisons, community reviews, and benchmark data to find the right fit for your workflow.
How much does Whisper.cpp cost?
Whisper.cpp is free-open-source. It is completely free and open source to self-host.
Pairs well with
Complementary tools, models, and hardware
Comments coming soon
Configure NEXT_PUBLIC_GISCUS_REPO_ID and NEXT_PUBLIC_GISCUS_CATEGORY_ID at giscus.app to enable.