# CrispASR: One C++ binary for multilingual ASR and TTS models > Source: > Published: 2026-08-31 22:22:23+00:00 **One C++ binary, 54 ASR backends + 55 TTS engines + multilingual text translation, zero Python dependencies.** CrispASR started as a fork of [whisper.cpp](https://github.com/ggml-org/whisper.cpp) and extends that base into a **unified speech engine** called `crispasr` , backed by full ggml C++ runtimes for major open-weights ASR *and* TTS architectures. One build, one binary, one consistent CLI — pick the backend at the command line or let CrispASR auto-detect it from your GGUF file. See [Text-to-Speech](#text-to-speech-models) for the TTS side. ``` bash $ crispasr -m ggml-base.en.bin -f samples/jfk.wav # OpenAI Whisper $ crispasr -m parakeet-tdt-0.6b.gguf -f samples/jfk.wav # NVIDIA Parakeet $ crispasr -m canary-1b-v2.gguf -f samples/jfk.wav # NVIDIA Canary $ crispasr -m voxtral-mini-3b-2507.gguf -f samples/jfk.wav # Mistral Voxtral $ crispasr --backend qwen3 -m auto -f samples/jfk.wav # -m auto downloads $ crispasr --backend kokoro -m auto --tts "Hello world" --tts-output out.wav # TTS ``` No Python. No PyTorch. No separate per-model binary. No `pip install` . Just one C++ binary and a GGUF file. **Browser**: All backends compile to WebAssembly (4.3 MB) via `build-wasm.sh` . Multithreaded, runs entirely client-side with COOP/COEP headers. **Demo**: [HuggingFace Space](https://huggingface.co/spaces/cstr/CrispASR) — live transcription + TTS + language detection, auto-deployed from `hf-space/` . | Project | What it does | |---|---| | This repo — C++ speech engine. 54 ASR + 55 TTS backends, CLI + HTTP server + C-ABI + Python/Rust/Dart/Go/Ruby/Java bindings. | | Cross-platform Flutter transcription app built on CrispASR. Desktop + mobile, model browser with download queue, mic capture, SRT/VTT/JSON export, diarization, batch processing. Fully offline. | | Text-related engine via ggml — same philosophy as CrispASR but for embeddings, retrieval, OCR and OMR, Math and Music Notation. Numerous architectures (XLM-R, Qwen3-Embed, Gemma3, ModernBERT, ...), dense + sparse + ColBERT + reranking. PP-OCR, Tesseract, EasyOCR, InternVL2, etc. Python/Rust/Dart bindings. | | Python ASR GUI with 9 backends (faster-whisper, mlx-whisper, voxtral, insanely-fast-whisper, ...). The Python counterpart to CrispASR's C++ approach. | — new to CrispASR? Two commands to your first working audio, no repo clone needed**Start here**[Supported backends](#supported-backends)—[ASR](#asr-backends)+[TTS](#text-to-speech-models)+[translation](#translation)+[post-processing](#post-processing-models)+[music & audio analysis](#music--audio-analysis)[Feature matrix](#feature-matrix)[Install & build](#install--build)— quick install (full guide in[docs/install.md](/CrispStrobe/CrispASR/blob/main/docs/install.md));— the[which prebuilt Linux tarball to download](/CrispStrobe/CrispASR/blob/main/docs/install.md#prebuilt-linux-tarballs--which-one-to-download-355)`-hip` /`-vulkan` builds require the matching GPU driver and do**not** fall back to CPU (the`-cuda` tarballs do, since v0.8.30)[Quick start — ASR](#quick-start)[Troubleshooting](/CrispStrobe/CrispASR/blob/main/docs/troubleshooting.md)— it printed the banner and stopped, reading the exit code,`--no-gpu` bisect, which Windows zip— 52 engines: Kokoro, Qwen3-TTS, VibeVoice, dots.tts, Orpheus, Chatterbox, IndexTTS, Irodori, VoxCPM2, CosyVoice3, CSM, Dia, Zonos, Bark, Piper, MeloTTS, and more**Text-to-Speech (TTS)**[Streaming & live transcription](/CrispStrobe/CrispASR/blob/main/docs/streaming.md)[Server mode (HTTP API)](/CrispStrobe/CrispASR/blob/main/docs/server.md)[Concurrency, parallelism & scaling](/CrispStrobe/CrispASR/blob/main/docs/concurrency.md)— how one transcription uses multiple cores, concurrent server requests (`--server-workers` ), bulk offline transcription, replicas behind a load balancer[CLI reference](/CrispStrobe/CrispASR/blob/main/docs/cli.md)— flags, VAD, CTC alignment, output formats, auto-download, audio formats[Environment variables](/CrispStrobe/CrispASR/blob/main/docs/environment-variables.md)— the`CRISPASR__` convention, global knobs, and every per-backend variable[Language bindings](/CrispStrobe/CrispASR/blob/main/docs/bindings.md)— Python / Rust / Dart / Go / Java / JavaScript / Ruby / mobile[Benchmarking CrispASR](/CrispStrobe/CrispASR/blob/main/docs/benchmarking.md)— how to measure transcribe time (not cold start): server/in-process reps, proof-of-work rules, phase-timing env vars[Architecture](/CrispStrobe/CrispASR/blob/main/docs/architecture.md)— layered layout,`src/core/` primitives, regression discipline[Contributing — adding a new backend](/CrispStrobe/CrispASR/blob/main/docs/contributing.md)— 5-file recipe, ground-truth diff workflow[Regression matrix](/CrispStrobe/CrispASR/blob/main/docs/regression-matrix.md)—`tools/test-all-backends.py` capability tiers— synthetic-audio marking (watermark + C2PA + spoken disclaimer), what counts as a voice clone, the speaker-biometrics boundary, why there is no emotion recognition, and what stays your duty as deployer**EU AI Act**[Quantize models](/CrispStrobe/CrispASR/blob/main/docs/quantize.md)—`crispasr-quantize` for all backends[GPU backend selection](#gpu-backend-selection)[Debugging & profiling](#debugging--profiling)[Credits](#credits) Everything below this section is a catalogue — 100+ backends, browse it when you need one. If you just want CrispASR *working*, this is the whole path. No repo clone, no Python, no model hunting. Download one file from [ Releases](https://github.com/CrispStrobe/CrispASR/releases/latest) and unzip it: | Platform | Download | Notes | |---|---|---| Windows | `crispasr-windows-x86_64-cpu.zip` | Needs AVX2 (2013+ Intel / 2015+ AMD). Older CPU → `…-cpu-legacy.zip` | Windows + NVIDIA | `crispasr-windows-x86_64-cuda.zip` | Self-contained; a CUDA Toolkit install is not required. CUDA-13-native build: `…-cuda13.zip` (Turing+) | macOS | `crispasr-macos.tar.gz` | Metal GPU support built in | Linux | `crispasr-linux-x86_64.tar.gz` | `…-cuda.tar.gz` / `…-vulkan.tar.gz` for GPU | Prefer to build it yourself? See [Install & build](#install--build). The `-hip` and `-vulkan` builds require the matching driver and do **not** fall back to CPU; the Linux `-cuda` tarballs do fall back. Check it runs — this should print a version banner and exit: ``` crispasr --version # Windows: .\crispasr.exe --version ``` `-m auto` downloads the model on first use (~135 MB here) and reuses it afterwards — nothing to find or install. It lands in `~/.cache/crispasr/` (`%USERPROFILE%\.cache\crispasr` on Windows). ``` crispasr --backend kokoro -m auto --tts "The quick brown fox jumps over the lazy dog." --tts-output hello.wav # crispasr: TTS output written to 'hello.wav' (78000 samples @ 24000 Hz, 3.25 sec) ``` Play `hello.wav` . That is the TTS half working. ``` crispasr --backend parakeet -m auto -f hello.wav -l en # crispasr: transcribed 3.2s audio in 0.32s (10.1x realtime) # The quick brown fox jumps over the lazy dog. ``` (~467 MB on first run. `-l en` skips language auto-detection, which would otherwise fetch a small extra model.) Both halves now work — swap in your own `.wav` and you are running. | You want to… | Go to | |---|---| | Clone a voice from a recording | `--i-have-rights` | [Which backend should I pick?](#which-backend-should-i-pick)[docs/cli.md](/CrispStrobe/CrispASR/blob/main/docs/cli.md)[docs/streaming.md](/CrispStrobe/CrispASR/blob/main/docs/streaming.md)[docs/server.md](/CrispStrobe/CrispASR/blob/main/docs/server.md)`crispasr --list-backends` Add `-v` to any command for verbose progress, and `--dry-run-resolve` to print which model files it would open (and whether they're on disk) without loading anything. If a command printed its banner and then simply stopped — no error, no output file — that is a crash, not a refusal, and the exit code identifies it in one step. See ** docs/troubleshooting.md**. CrispASR ships **54 ASR backends** for transcription/translation and **55 TTS engines** for synthesis. It also ships audio-to-audio S2S backends, including Sidon restoration and the VoxCPM2 AudioVAE speech upscaler; see the [feature matrix](/CrispStrobe/CrispASR/blob/main/docs/feature-matrix.md) for the complete capability list. Pick at the CLI with `--backend NAME` , or omit it to let the binary auto-detect from the GGUF metadata. Jump to the [TTS table](#text-to-speech-models) for the synthesis side. | Backend | Model | Architecture | Languages | License | |---|---|---|---|---| whisper | `ggml-base.en.bin` | **whisper**`distil-whisper/distil-large-v3` **parakeet**`nvidia/parakeet-tdt-0.6b-v3` **parakeet**`nvidia/parakeet-tdt-0.6b-v2` **parakeet**`nvidia/parakeet-tdt-1.1b` **parakeet**`nvidia/parakeet-tdt_ctc-110m` **parakeet**`nvidia/parakeet-tdt_ctc-1.1b` **parakeet**`nvidia/parakeet-tdt_ctc-0.6b-ja` **reazonspeech**`reazon-research/reazonspeech-nemo-v2` **fastconformer-ctc**`nvidia/parakeet-ctc-0.6b` **fastconformer-ctc**`nvidia/parakeet-ctc-1.1b` **fastconformer-ctc**`grider-transwithai/parakeet-ctc-1.1b-ja` **canary**`nvidia/canary-1b-v2` `-sl/-tl` )**canary-qwen**`nvidia/canary-qwen-2.5b` **lfm2-audio**`LiquidAI/LFM2.5-Audio-1.5B` **lfm2-audio**`LiquidAI/LFM2.5-Audio-1.5B-JP` **mini-omni2**`gpt-omni/mini-omni2` **cohere**`CohereLabs/cohere-transcribe-03-2026` **cohere**`efwkjn/cohere-asr-ja-v0.1` **granite**[,](https://huggingface.co/ibm-granite/granite-speech-3.3-2b)`ibm-granite/granite-speech-{3.2-8b,3.3-2b,3.3-8b}` `granite-4.0-1b-speech` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#granite--granite-41--granite-41-plus--granite-41-nar))**granite-4.1**`ibm-granite/granite-speech-4.1-2b` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#granite--granite-41--granite-41-plus--granite-41-nar))**granite-4.1-plus**`ibm-granite/granite-speech-4.1-2b-plus` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#granite--granite-41--granite-41-plus--granite-41-nar))**granite-4.1-nar**`ibm-granite/granite-speech-4.1-2b-nar` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#granite--granite-41--granite-41-plus--granite-41-nar))**fastconformer-ctc**`nvidia/stt_en_fastconformer_ctc_large` **voxtral**`mistralai/Voxtral-Mini-3B-2507` **voxtral4b**`mistralai/Voxtral-Mini-4B-Realtime-2602` **qwen3**`Qwen/Qwen3-ASR-0.6B` **qwen3-1.7b**`Qwen/Qwen3-ASR-1.7B` **qwen3-ja-anime**`jaykwok/Qwen3-ASR-1.7B-JA-Anime-Galgame-hf` **mega-asr**`zhifeixie/Mega-ASR` **higgs-stt**`bosonai/higgs-audio-v3-stt` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#higgs-stt))**wav2vec2**`jonatasgrosman/wav2vec2-large-xlsr-53-english` **wav2vec2**`facebook/data2vec-audio-base-960h` **wav2vec2**`facebook/hubert-large-ls960-ft` **glm-asr**`zai-org/GLM-ASR-Nano-2512` **kyutai-stt**`kyutai/stt-1b-en_fr` **kyutai-stt**`kyutai/stt-2.6b-en` **firered-asr**`FireRedTeam/FireRedASR2-AED` **moonshine**`UsefulSensors/moonshine-{tiny,base}` **moonshine‑de**`fidoriel/moonshine-base-de` **moonshine‑tiny‑de**`fidoriel/moonshine-tiny-de` **moonshine-streaming**`UsefulSensors/moonshine-streaming-{tiny,small,medium}` **gemma4-e2b**`google/gemma-4-E2B-it` **gemma4-e4b**`google/gemma-4-E4B-it` `--backend gemma4-e2b` **omniasr**`omniASR-CTC-1B-v2` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#omniasr-ctc--llm--unlimited))**1600+****omniasr‑300m**`omniASR-CTC-300M-v2` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#omniasr-ctc--llm--unlimited))**1600+****omniasr-llm**`omniASR-LLM-300M-v2` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#omniasr-ctc--llm--unlimited))**1600+****omniasr-llm**`omniASR-LLM-Unlimited-300M-v2` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#omniasr-ctc--llm--unlimited))**1600+****vibevoice**`microsoft/VibeVoice-ASR` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#vibevoice))**vibevoice-bitnet**`VibeVoice-ASR-BitNet` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#vibevoice))**mimo-asr**`XiaomiMiMo/MiMo-V2.5-ASR` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#mimo-asr))**ark-asr*** experimental/WIP*[(base](https://huggingface.co/cstr/ark-asr-3b-GGUF)`cstr/ark-asr-3b-GGUF` [)](https://huggingface.co/AutoArk-AI/ARK-ASR-3B)`AutoArk-AI/ARK-ASR-3B` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#ark-asr))**moss-audio**`OpenMOSS-Team/MOSS-Audio-4B-Instruct` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#moss-audio))**moss-transcribe**`OpenMOSS-Team/MOSS-Transcribe-preview-2B` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#moss-transcribe))**moss-diarize**`OpenMOSS-Team/MOSS-Transcribe-Diarize-0.9B` **whisper***(tiron)** experimental*[(base](https://huggingface.co/cstr/tiron-GGML)`Trelis/tiron` [)](https://huggingface.co/Trelis/tiron)`Trelis/tiron` **funasr**`FunAudioLLM/Fun-ASR-Nano-2512` **fun-asr-mlt-nano**`FunAudioLLM/Fun-ASR-MLT-Nano-2512` **paraformer**`funasr/paraformer-zh` **foxnose***(speaker diarization)*`Wespeaker/wespeaker-voxceleb-resnet34-LM` `--diarize-method foxnose` : WeSpeaker ResNet34-LM 256-d embeddings + GMM/BIC speaker counting + spectral clustering + Viterbi temporal smoothing ([more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#foxnose-diarize)). 3.18 % DER on VoxConverse dev vs the upstream reference implementation's 3.07 %**gigaam**[(base](https://huggingface.co/cstr/gigaam-v3-GGUF)`ai-sage/GigaAM-v3` [)](https://huggingface.co/ai-sage/GigaAM-v3)`ai-sage/GigaAM-v3` `e2e_rnnt` / `e2e_ctc` emit punctuation + casing + ITN from a SentencePiece vocab, `rnnt` / `ctc` emit bare lowercase Cyrillic ([more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#gigaam))**sensevoice**`FunAudioLLM/SenseVoiceSmall` `-oj` JSON expose the tags as separate fields. Upstream's emotion classifier is **not exposed**— see[EU AI Act](/CrispStrobe/CrispASR/blob/main/docs/eu-ai-act.md#41-emotion-recognition--removed-not-gated)| Backend | Model | Architecture | Input / output | License | |---|---|---|---|---| sidon | `KevinAHM/Sidon-GGUF` | [)](https://huggingface.co/sarulab-speech/sidon-v0.1) `sarulab-speech/sidon-v0.1` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#sidon))**voxcpm2-vae**[, converted with](https://huggingface.co/openbmb/VoxCPM2)`openbmb/VoxCPM2` `--vae-only` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#voxcpm2-vae)) ``` huggingface-cli download KevinAHM/Sidon-GGUF sidon-v0.1-f16.gguf --local-dir models crispasr -m models/sidon-v0.1-f16.gguf -f input.wav --s2s --s2s-output restored.wav python models/convert-voxcpm2-to-gguf.py --input openbmb/VoxCPM2 \ --output models/voxcpm2-vae-f32.gguf --vae-only crispasr -m models/voxcpm2-vae-f32.gguf -f input.wav --s2s \ --s2s-output upscaled.wav ``` Synthesis backends, driven by the `--tts` flag and a `--tts-output PATH.wav` . See the dedicated [Text-to-Speech](#text-to-speech-models) section below for quick-start commands and engine selection guidance. | Backend | Models | Architecture | Languages | License | |---|---|---|---|---| miotts | `MioTTS-0.6B` | **vibevoice-tts**[,](https://huggingface.co/cstr/vibevoice-realtime-0.5b-GGUF)`VibeVoice-Realtime-0.5B` `VibeVoice-1.5B` **kugelaudio**`kugelaudio-0-open` **qwen3-tts**[,](https://huggingface.co/cstr/qwen3-tts-0.6b-base-GGUF)`Qwen3-TTS-12Hz-0.6B-Base` [,](https://huggingface.co/cstr/qwen3-tts-1.7b-base-GGUF)`1.7B-Base` `1.7B-VoiceDesign` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#qwen3-tts))**qwen3-tts-customvoice**`1.7B-CustomVoice` `--voice ` ); optional style via `--instruct` (e.g. "spoke very slowly") ([more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#qwen3-tts))**moss-tts**`OpenMOSS-Team/MOSS-TTS-v1.5` `--voice ref.wav` ; `--backend moss-tts -m --codec-model ` **moss-tts-local**`OpenMOSS-Team/MOSS-TTS-Local-Transformer-v1.5` `--backend moss-tts-local -m --codec-model ` **omnivoice**`k2-fsa/OmniVoice` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#omnivoice))**melotts**[EN_V2](https://github.com/myshell-ai/MeloTTS)`myshell-ai/MeloTTS` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#melotts))**piper**[community voices](https://github.com/rhasspy/piper)`rhasspy/piper` `--g2p-dict` )**kokoro**[+ German backbones](https://huggingface.co/hexgrad/Kokoro-82M)`hexgrad/Kokoro-82M` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#kokoro))**orpheus**[+](https://huggingface.co/cstr/orpheus-3b-0.1-ft-GGUF)`Orpheus-3B-FT` `SNAC 24 kHz` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#orpheus))**chatterbox**[+ Nano/turbo/fine-tune variants](https://huggingface.co/cstr/chatterbox-GGUF)`cstr/chatterbox-GGUF` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#chatterbox--chatterbox-turbo--chatterbox-nano--chatterbox-finnish-nano--kartoffelbox-turbo--lahgtna-chatterbox))`chatterbox-finnish-nano` ) fine-tunes**indextts**`cstr/indextts-1.5-GGUF` **voxcpm2-tts**`cstr/voxcpm2-GGUF` `--voice ` **voxtral-tts**`mistralai/Voxtral-4B-TTS-2603` **cosyvoice3-tts**`cstr/cosyvoice3-0.5b-2512-GGUF` `--voice ` , or any WAV via `--voice ref.wav --ref-text ""` . `--backend cosyvoice3-tts-rl` selects upstream's RL-tuned talker (same companions)**csm**`cstr/csm-1b-GGUF` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#csm))**lfm2-audio**[+](https://huggingface.co/cstr/lfm2-audio-1.5b-GGUF)`cstr/lfm2-audio-1.5b-GGUF` `jp` **dia**`nari-labs/Dia-1.6B` `[S1]` /`[S2]` tags (use >100-char prompts)**zonos-tts**[+](https://huggingface.co/cstr/zonos-v0.1-transformer-GGUF)`cstr/zonos-v0.1-transformer-GGUF` `cstr/dac-44khz-GGUF` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#zonos-tts))**bark**`cstr/bark-small-GGUF` `.npz` prompts (`--voice ` )**speecht5**`cstr/speecht5-tts-GGUF` `--voice ` )**fastpitch**`cstr/fastpitch-en-GGUF` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#fastpitch))**bananamind-tts**`Banaxi-Tech/BananaMind-TTS-V2.1-Preview` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#bananamind-tts))**parler-tts**`cstr/parler-tts-mini-v1.1-GGUF` `--instruct` )**outetts**`cstr/outetts-0.3-1b-GGUF` `--voice ` )**pocket-tts**`cstr/pocket-tts-GGUF` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#pocket-tts))**tada**[+](https://huggingface.co/cstr/tada-tts-1b-GGUF)`cstr/tada-tts-1b-GGUF` `HumeAI/tada-3b-ml` `tada-ref.gguf` , custom voices via `--voice ` built with `models/convert-tada-ref-to-gguf.py` ([more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#tada))**TTS feature matrix** | Backend | Voice cloning | Sampling | kHz | Auto-download | Flash attn | |---|---|---|---|---|---| | vibevoice-tts | yes | temp | 24 | yes | yes | | qwen3-tts | yes* | temp | 24 | yes | yes | | omnivoice | yes | temp | 24 | — | — | | kokoro | — | — | 24 | yes | — | | orpheus | — | temp | 24 | yes | yes | | chatterbox | yes | temp | 24 | yes | yes | | outetts | yes (JSON) | temp | 24 | yes | yes | | indextts | yes | temp | 24 | yes | yes | | voxcpm2-tts | yes | — | 48 | yes | — | | cosyvoice3-tts | yes | temp | 24 | yes | yes | | f5-tts | yes | — | 24 | yes | — | | irodori-tts | yes (WAV) | VoiceDesign: `--instruct` | 48 | yes | — | | csm | — | temp | 24 | yes | — | | dia | — | temp | 44 | yes | — | | bark | yes (.npz) | temp | 24 | yes | — | | speecht5 | yes (xvec) | — | 16 | yes | — | | parler-tts | — | temp | 44 | yes | — | | fastpitch | — | — | 22 | — | — | | piper | — | — | 22 | — | — | | pocket-tts | yes | temp | 24 | yes | — | | tada | yes | temp | 24 | yes | — | | dots-tts | yes (`--voice ref.wav` ) | 16-step CFG Euler | 48 | yes | — | | confucius4-tts | yes (`--voice ref.wav` ) | 25-step CFG Euler | 22.05 | yes | — | * CustomVoice variant only; Base uses baked speakers via `--voice ` . **Output language.** `-tl ` (or `-l` ) selects the language to speak; `cosyvoice3-tts` , `qwen3-tts` and `moss-tts` act on it natively. For cross-lingual **cloning** — an English reference clip speaking German, the subtitle-dubbing case — also pass `-sl ` for the language the reference is spoken in, so cosyvoice3 drops the reference transcript instead of carrying its accent. Over HTTP: `"language"` + `"source_lang"` on `POST /v1/audio/speech` . See [ docs/tts.md](/CrispStrobe/CrispASR/blob/main/docs/tts.md#output-language-and-cross-lingual-cloning--tl---sl). Text-to-text translation, distinct from the audio-side `--translate` flag (which routes audio → English text on whisper / canary / etc.). Driven by `--text "..." -sl -tl ` . | Backend | Models | Architecture | Languages | License | |---|---|---|---|---| m2m100 | `facebook/m2m100_418M` | [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#m2m100--wmt21))**m2m100-wmt21**[+](https://huggingface.co/cstr/wmt21-dense-24-wide-en-x-GGUF)`facebook/wmt21-dense-24-wide-en-x` `facebook/wmt21-dense-24-wide-x-en` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#m2m100--wmt21))`en-x` / `x-en` checkpoints)**madlad**`google/madlad400-3b-mt` [more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#madlad)) ``` # m2m100 base (production-ready) ./build/bin/crispasr --backend m2m100 -m auto \ --text "Hello world, how are you today?" \ -sl en -tl de # → Hallo Welt, wie bist du heute? # WMT21 dense (English ↔ X, 4.7B — auto-downloads ~2.5 GB). # Two separate checkpoints: en-x for English-source, x-en for # English-target. Pick the one matching your `-sl`/`-tl` direction # (or pass an explicit `-m ` to load the other manually). ./build/bin/crispasr --backend m2m100-wmt21 -m auto \ --text "The president said he would not attend." \ -sl en -tl de # uses wmt21-dense-24-wide-en-x ./build/bin/crispasr --backend m2m100-wmt21 \ -m models/wmt21-dense-24-wide-x-en-q4_k.gguf \ --text "Le président a dit qu'il ne serait pas présent." \ -sl fr -tl en # uses wmt21-dense-24-wide-x-en # MADLAD-400 3B (419 languages, bit-token-identical to Python SP) ./build/bin/crispasr --backend madlad -m auto \ --text "Hello world." \ -sl en -tl ta ``` For 2-stage pipelines (e.g., ASR → m2m100), use the dedicated `--tr-sl` / `--tr-tl` flags; they fall back to `-sl` / `-tl` when unset, so single-stage standalone usage is just `-sl/-tl` . Work with all backends. | Model | Task | Architecture | Languages | License | HuggingFace | |---|---|---|---|---|---| FireRedPunc | Punctuation restoration | BERT-base (12L, d=768), 5 classes | Chinese + English | Apache-2.0 | `cstr/fireredpunc-GGUF` | **fullstop-punc**`cstr/fullstop-punc-multilang-GGUF` **punctuate-all**`cstr/punctuate-all-GGUF` **PCS**`--punc-model pcs` **truecaser‑lstm**`--truecase-model lstm` **truecaser‑crf**`--truecase-model crf` **truecaser‑de**`--truecase-model auto` **CLD3**`cstr/cld3-GGUF` **GlotLID-V3**`cstr/glotlid-GGUF` **LID-176**`cstr/fasttext-lid176-GGUF` Shared codec modules used by TTS backends. Also available standalone for encode/decode. | Model | Architecture | Sample Rate | Token Rate | License | HuggingFace | |---|---|---|---|---|---| MioCodec v2 | WavLM encoder → FSQ(12800) → Transformer decoder + AdaLN-Zero + SnakeBeta upsampler + iSTFT | 44.1 kHz | 25 Hz (341 bps) | MIT | `cstr/miocodec-v2-44k-GGUF` | **SNAC 24 kHz**`cstr/snac-24khz-GGUF` All runtimes share ggml-based inference. The speech-LLM backends (**qwen3**, **voxtral**, **voxtral4b**, **granite**, **glm-asr**, **kyutai-stt**) inject audio encoder frames directly into an autoregressive language model's input embeddings, instead of using a dedicated CTC/transducer/seq2seq decoder. The **fastconformer-ctc** backend hosts the NeMo FastConformer-CTC standalone ASR family — `stt_en_fastconformer_ctc_{large,xlarge,xxlarge}` and the architecturally-identical `parakeet-ctc-{0.6b,1.1b}` (different training data + tokenizer, same encoder + head shape) — with greedy CTC decoding. Same C++ runtime as the canary-ctc aligner. Beyond speech, CrispASR runs several music/audio analysis tasks — each a small GGUF with the architecture auto-detected, no Python. See [ docs/cli.md](/CrispStrobe/CrispASR/blob/main/docs/cli.md) for the per-task flags and output formats. **Source separation**(`--separate` ) — split a mix into stems (`_.wav` ) via**mel-band-roformer**(vocal/instrumental, MIT) or** htdemucs**(4-stem).`--stems vocals,drums` selects a subset;`--sep-output-dir` sets the output location.**Piano transcription**(`--backend piano-transcription` ) — piano audio → MIDI note events (88 keys @ 100 fps, ByteDance/Kong CRNN; F16 GGUF ≈ 77 MB).**Guitar tablature**(`--tab` ) — per-frame fret-per-string grid via**TabCNN**(Wiggins & Kim, ISMIR 2019; CC BY 4.0 weights). The backend emits per-string emission scores, not a decided tablature — run your own constrained Viterbi via`crispasr_session_tab_emissions()` for playable output.**Beat / downbeat tracking**(`--beats` ) — beat grid via**Beat This!**(CPJKU, ISMIR 2024; MIT for code*and*weights, no patent-encumbered DBN).**Chord recognition**(`--chords` ) — chord timeline (`.lab` ) via**BTC**(ISMIR 2019). Weights are CC-BY-NC-SA, gated behind`--accept-license cc-by-nc-sa-4.0` .**Pitch / F0 estimation**(`--pitch` ) — monophonic pitch track via**CREPE**(MIT). Run `crispasr --list-backends` to see it live. Each backend declares capabilities at runtime; if you ask for a feature the selected backend does not support, CrispASR prints a warning and silently ignores the flag. **Sortable / filterable view:** [ docs/feature-matrix.html](/CrispStrobe/CrispASR/blob/main/docs/feature-matrix.html) — click any column header to sort, type to filter rows, click cap pills to require a capability. Generated from `crispasr --list-backends-json` (single source of truth — drift impossible). Regenerate via `python tools/gen-feature-matrix.py` . A Markdown twin lives at [.](/CrispStrobe/CrispASR/blob/main/docs/feature-matrix.md) `docs/feature-matrix.md` The static table below is a curated subset focusing on the ASR backends and the cross-cutting features that matter for ASR pipelines. The full 109-backend × 27-cap surface is in the generated views. | Feature | whisper | parakeet | canary | cohere | granite | granite‑4.1 | voxtral | voxtral4b | qwen3 | fc‑ctc | wav2vec2 | glm‑asr | kyutai‑stt | firered | moonshine | moon‑stream | omniasr | omniasr‑llm | vibevoice | gemma4‑e2b | mimo‑asr | funasr | paraformer | sensevoice | |---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---| | Native timestamps | ✔ | ✔ | ✔ | ✔ | ✔ | ||||||||||||||||||| | CTC timestamps | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||| | Word-level timing | ✔ | ✔ | ✔ | ✔ | `-am` | ✔† | `-am` | `-am` | `-am` | `-am` | `-am` | `-am` | ✔ | `-am` | `-am` | `-am` | `-am` | `-am` | `-am` | `-am` | `-am` | `-am` | || | Per-token confidence | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||| | Language auto-detect | ✔ | ✔ | LID | LID | LID | LID | LID | LID | ✔ | LID | LID | ✔ | LID | LID | LID | LID | LID | LID | LID | ✔ | LID | LID | LID | ✔ | | Speech translation | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||||||||||||||||| | Speaker diarization | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | | Grammar (GBNF) | ✔ | ||||||||||||||||||||||| | Temperature sampling | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ||||| | Beam search | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ||||||||||||| | Flash attention | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | || | Punctuation toggle | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ||||||||| | Punc restoration | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | pp | | Source / target language | ✔ | ✔ | ✔ | ✔ | ✔ | ||||||||||||||||||| Audio Q&A (`--ask` ) | * | * | ✔ | * | * | * | * | ||||||||||||||||| | Streaming | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | Auto-download (`-m auto` ) | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | KV quant (`CRISPASR_KV_QUANT` , plus per-half `_K` / `_V` ) | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | |||||||||||||| mmap weights (`CRISPASR_GGUF_MMAP` ) | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | ✔ | | | TTS | ✔ | The matrix above covers 24 ASR backends. **Additional ASR backends** not shown: `nemotron` (39-lang streaming ASR with cache-aware FastConformer + RNN-T), `lfm2-audio` (ASR + TTS + S2S in one model), `moss-audio` (audio understanding + ASR), `moss-transcribe` (Qwen3-Omni encoder + Qwen3-1.7B ASR), `mini-omni2` (ASR + TTS + S2S), `kugelaudio` (7B audio understanding). See [ docs/feature-matrix.md](/CrispStrobe/CrispASR/blob/main/docs/feature-matrix.md) for the full 109-backend matrix. **TTS-only backends**( `kokoro` , `qwen3-tts` + variants, `vibevoice-tts` , `orpheus` + DE variants, `chatterbox` / `chatterbox-turbo` / `chatterbox-nano` / `kartoffelbox-turbo` / `lahgtna-chatterbox` , `dia` , `bark` , `outetts` , `zonos` , `csm` , `f5-tts` , `irodori-tts` , `parler-tts` , `speecht5` , `piper` , `fastpitch` , `pocket-tts` , `melotts` , `cosyvoice3` , `voxcpm2` , `tada-tts` ) all carry the TTS, AUTO_DOWNLOAD, TEMPERATURE, and FLASH_ATTN caps; per-backend cloning + voice-pack support is documented in the [Text-to-Speech models](#text-to-speech-models)table above and [. The vibevoice and lfm2-audio columns mark dual-mode (ASR + TTS) backends.](/CrispStrobe/CrispASR/blob/main/docs/tts.md) `docs/tts.md` **Key:** ✔ = native/built-in, `-am` = via CTC forced aligner (`-am canary-ctc-aligner.gguf` or `-am qwen3-forced-aligner.gguf` ), **LID** = via external language identification pre-step (`-l auto` ), **pp** = via `--punc-model` post-processor (FireRedPunc or fullstop-punc), * = experimental or partial support, † = PLUS variant only (native `[T:N]` word timestamps with `-owts` ; base uses `-am` ). granite-4.1 covers both the regular and `-plus` variants; granite-4.1-nar is a non-autoregressive variant with encoder+projector only (no LLM decode features). The **KV quant** row marks backends that honor `CRISPASR_KV_QUANT={f16,q8_0,q4_0}` — CTC-style backends without a KV cache (parakeet, fc-ctc, wav2vec2, kyutai-stt, firered, moonshine variants, omniasr-CTC) don't apply. The same backends also honor the per-half `CRISPASR_KV_QUANT_K` / `CRISPASR_KV_QUANT_V` overrides (llama.cpp `--cache-type-k` / `--cache-type-v` parity) for asymmetric K-vs-V precision; common recipe `K=q8_0 V=q4_0` saves ~40 % more KV memory than symmetric Q8_0. The **mmap weights** row marks backends consuming `core_gguf::load_weights()` and therefore honoring `CRISPASR_GGUF_MMAP=1` ; whisper itself uses upstream's loader and is unaffected. See [ docs/cli.md](/CrispStrobe/CrispASR/blob/main/docs/cli.md) Memory footprint for usage + recommended combos. **Speaker diarization** as a post-processing step via `--diarize` : `energy` /`xcorr` — stereo-only, no extra deps`foxnose` —**best accuracy, no external deps**: WeSpeaker ResNet34-LM embeddings + GMM/BIC speaker counting + spectral clustering + Viterbi smoothing. Estimates the speaker count rather than needing it up front;`--diarize-embedder auto` fetches the GGUF (24 MB, CC-BY-4.0). 7.3 % DER on VoxConverse dev where`pyannote` + TitaNet scores 7.8 %, and 3.18 % vs the upstream reference's 3.07 % when scored on turns ([more](/CrispStrobe/CrispASR/blob/main/docs/architecture.md#foxnose-diarize))`pyannote` — native GGUF (no Python, no sherpa-onnx); add`--diarize-embedder auto` (TitaNet) or`--diarize-embedder indextts` (ECAPA-TDNN) for globally stable speaker IDs across long files`sherpa` /`ecapa` — external[sherpa-onnx](https://github.com/k2-fsa/sherpa-onnx)subprocess; runs once globally on full audio for consistent speaker IDs (#110)`vad-turns` — mono-friendly gap-based proxy The server endpoint supports `response_format=diarized_json` for structured speaker-labelled output with normalised speaker letters (A, B, C …) — see [ docs/server.md](/CrispStrobe/CrispASR/blob/main/docs/server.md#diarized-json-format-206). Full reference + tuning knobs (cluster threshold, max speakers, pluggable embedder adapters): see [ docs/cli.md#diarization](/CrispStrobe/CrispASR/blob/main/docs/cli.md#diarization). **Language identification** for backends without native LID: `--lid-backend whisper` (default, 75 MB ggml-tiny.bin), `--lid-backend silero` (native GGUF, 16 MB, 95 languages), or `--lid-backend firered` (FireRedLID, 1.7 GB, 120 languages — Conformer encoder + Transformer decoder). **Voice activity detection**: `--vad` uses the default Silero VAD (~885 KB, auto-downloaded). Each VAD segment is transcribed independently, producing separate SRT/VTT entries with correct timestamps. Use `--vad --split-on-punct` for best subtitle output. Four VAD backends: Silero (default), FireRedVAD (`-vm firered` , recommended), MarbleNet (`-vm marblenet` , 439 KB, 6 languages), Whisper-VAD-EncDec (`-vm whisper-vad` , experimental). **Punctuation restoration** (`--punc-model` ): CTC-based backends output lowercase without punctuation. Named shortcuts: `auto` /`firered` (Chinese+English), `fullstop` (EN/DE/FR/IT, XLM-R-large), `punctuate-all` (12 languages, XLM-R-base), `pcs` (47 languages, punc + truecasing + sentence boundary detection in one model). Or pass a GGUF path directly. Also available via Python/Rust/Dart wrappers (`crispasr.PuncModel` ). **Truecasing** (`--truecase-model` ): Restore German noun/name capitalization in lowercase ASR output. Three options in ascending quality: `auto` (statistical, 1.7 MB), `crf` (CRF with context, 8.5 MB), `lstm` (BiLSTM char-level, 3.2 MB, **recommended** — 97.9% F1, handles adjective/noun distinction and formal "Ihnen"). All auto-download from [ cstr/truecaser-de](https://huggingface.co/cstr/truecaser-de). Or use `--punc-model pcs` for neural punc + truecasing in one pass (47 languages).## Which backends produce punctuation natively? | Backend | Punctuation | Capitalization | Notes | |---|---|---|---| | whisper | ✔ | ✔ | Full punctuation and casing | | parakeet | ✔ | ✔ | | | canary | ✔ | ✔ | | | cohere | ✔ | ✔ | Toggleable via `--no-punctuation` | | granite | ✔ | ✔ | LLM output | | voxtral | ✔ | ✔ | LLM output | | voxtral4b | ✔ | ✔ | LLM output | | qwen3 | ✔ | ✔ | LLM output | | funasr | ✔ | ✔ | LLM output (Qwen3-0.6B decoder). Chinese chars carry full-width period; mlt-nano variant adds Latin-script casing + punctuation. | | sensevoice | ✔ | ✔ | CTC output with native ITN — toggle off via `--no-punctuation` , which controls Arabic-digit vs spelled-out numerals + comma/period emission. | | paraformer | no | no | NAR character-level output — add `--punc-model` | | gigaam | ✔ (`e2e_*` ) | ✔ (`e2e_*` ) | The `e2e_*` revisions carry punctuation + casing + inverse text normalization in the SentencePiece vocabulary. The charwise `ctc` / `rnnt` revisions emit lowercase Cyrillic with no punctuation — but auto-restoration is still suppressed for them, because the auto-enabled FireRedPunc is a Chinese/English model and injects full-width CJK punctuation into Russian. Use an `e2e_*` revision for punctuated output, or pass an explicit `--punc-model` . | | glm-asr | ✔ | ✔ | LLM output | | kyutai-stt | ✔ | ✔ | LLM output | | moonshine | ✔ | ✔ | Encoder-decoder output | fastconformer-ctc | no | no | CTC — add `--punc-model` | wav2vec2 | no | no | CTC — add `--punc-model` | firered-asr | no | no | CTC — add `--punc-model` | omniasr (CTC) | no | no | CTC — add `--punc-model` | omniasr (LLM) | ✔ | ✔ | Autoregressive decoder | Other freely-licensed alternatives that could be added: [felflare/bert-restore-punctuation](https://huggingface.co/felflare/bert-restore-punctuation) (MIT, English, includes truecasing), [xashru/punctuation-restoration](https://github.com/xashru/punctuation-restoration) (Apache-2.0, 40+ languages, BiLSTM-CRF). **Progressive subtitle output** (`--flush-after` ): By default, non-whisper backends buffer all segments and print output at the end. For real-time subtitle consumption (PotPlayer, custom media players), use `--flush-after 1` to print each SRT entry to stdout immediately after its VAD segment is transcribed: ``` crispasr --backend parakeet -m parakeet.gguf --vad --flush-after 1 -osrt -f long_audio.wav # SRT entries appear progressively as each segment finishes ``` **JSON output with language detection**: When using `-l auto -oj` , the JSON output includes detected language info: ``` { "crispasr": { "backend": "cohere", "language": "en", "language_detected": "en", "language_confidence": 0.977, "language_source": "ecapa" }, "transcription": [...] } ``` | Need | Pick | |---|---| | Battle-tested, all features exposed | whisper | | Lowest English WER | cohere | Fastest (16x realtime on CPU) | moonshine (tiny), fc-ctc (10x) | | Multilingual + word timestamps + fast | parakeet (2.9x RT) | Multilingual with explicit language control | canary | Speech translation (X→en or en→X) | canary, voxtral, qwen3 | 30 languages + Chinese dialects | qwen3 | 1600+ languages | omniasr (CTC or LLM) | Realtime streaming ASR (native incremental encoder, ~2× RT feed; sub-second-token target deferred to phase 2) | voxtral4b | | Highest-quality offline speech-LLM | voxtral | | Apache-licensed speech-LLM | granite, voxtral, qwen3, omniasr-llm | Lightweight CTC-only (fast, no decoder) | wav2vec2, fc-ctc, data2vec, omniasr | Russian | gigaam (`e2e_rnnt` — 8.4 % avg WER, punctuation + ITN), whisper, qwen3 | Mandarin + Chinese dialects | firered-asr, qwen3, glm-asr, funasr, paraformer, sensevoice | Multilingual (31 langs) speech-LLM | fun-asr-mlt-nano, qwen3, omniasr-llm, gemma4-e2b | Multilingual (50+ langs) + LID + audio-event in one pass | sensevoice (encoder-only CTC, non-AR, 15× faster than Whisper-Large) | Audio-LLM backends (`qwen3` , `voxtral` , `granite` , `glm-asr` , etc.) run full transformer decoder stacks (28+ layers, 2048-dim) and are **dramatically slower on CPU** than encoder-only backends. On older dual-core hardware they can drop below 0.01× realtime. If you're on CPU-only hardware: - Prefer **moonshine**(16× RT),** fc-ctc**(10× RT),** parakeet**(2.9× RT), or** whisper**for usable speeds. - Use `--flush-after 1` to see results as each VAD slice completes instead of waiting for the entire file. - Use `-pp` /`--print-progress` for per-slice progress indicators on all backends (unified backends show slice-level progress; whisper shows encoder-level progress). - Quantize models to Q4_K or Q5_K to reduce memory and compute. Cohere, canary, granite, voxtral and voxtral4b need an explicit language code up front. If you don't know the language, pass `-l auto` and crispasr runs an optional LID pre-step before the main transcribe() call: ``` # Downloads ggml-tiny.bin (75 MB, 99 languages) on first use crispasr --backend cohere -m $TC/cohere-transcribe-q5_0.gguf \ -f unknown.wav -l auto # crispasr[lid]: detected 'en' (p=0.977) via whisper-tiny # crispasr: LID -> language = 'en' (whisper, p=0.977) ``` These LID providers are available: - `--lid-backend whisper` (default) — uses a small multilingual ggml-*.bin model via the crispasr C API. Auto-downloads ~75 MB on first use. 99 languages. - `--lid-backend silero` — native GGUF port of Silero's 95-language classifier. 16 MB F32. Runs as a ggml graph (multi-threaded SIMD on CPU, GPU offload on Metal/CUDA; on Vulkan the graph is routed to CPU pending an upstream kernel fix). Analyzes the first 30 s of audio (`CRISPASR_SILERO_LID_MAX_S` overrides);`CRISPASR_SILERO_LID_LEGACY=1` restores the old scalar path. - `--lid-backend ecapa` —**recommended**: ECAPA-TDNN (Apache-2.0). Purpose-built for language ID. Very high accuracy on TTS benchmark. Two variants via`--lid-model` :— VoxLingua107, 43 MB F16, 107 languages, ISO codes (en, de, ...).`cstr/ecapa-lid-107-GGUF` **Default.**— CommonLanguage, 40 MB F16, 45 languages, full names (English, German, ...).`cstr/ecapa-lid-commonlanguage-GGUF` - `--lid-backend firered` — FireRedLID (Conformer encoder + Transformer decoder). Q4_K (544 MB), 120 languages including Chinese dialects. Slower but covers more languages. - `--lid-backend probe` — no second model at all: ask the**ASR model itself**. Transcribes a 20 s clip once per language the model declares and keeps the best-scoring candidate (length × text-LID agreement × distinct-token ratio², the last term catching the repetitive output a wrong-language prompt produces). Currently implemented by**cohere**. The reason to prefer it is correctness, not just the saved download: an external detector knows 99 languages while Cohere Transcribe accepts 14 — and its Arabic finetune only`en` /`ar` — so external LID regularly returns a language the model was never trained on, and Cohere answers a wrong language*fluently*rather than failing. The probe cannot. Cost is one encode + one short decode per candidate, so it runs automatically only for a model with ≤ 4 languages (`CRISPASR_COHERE_PROBE_MAX_LANGS` );`CRISPASR_COHERE_PROBE_TEXTLID=0` drops the text-LID agreement term.**The ceiling is about cost, not accuracy.** Measured on the real models: the two-language Arabic finetune picks`ar` for an Arabic clip (p=0.675) and`en` for`samples/jfk.wav` (p=0.647); the 14-language base model, probed across all 14, also gets both right (`en` p=0.169,`ar` p=0.254) — it is simply slower than an external detector. The encoder output is language-independent, so the probe encodes**once** and decodes per candidate (encode is ~87 % of a pass); a 14-candidate probe measured**12 s → 4-5 s** against one-encode-per-candidate, byte-identical output.`CRISPASR_COHERE_PROBE_REUSE_ENC=0` restores the naive path.The one soft spot worth knowing: asking the model for a language it was *not*trained on can yield a clean translation rather than garbage, which a text LID then confirms — "fluent French out" is not evidence of French in. That is what makes a mismatched whitelist dangerous, and it is measurable: force a 14-language list onto the two-language Arabic finetune and its`fr` probe returns real French and wins. The real base model's`fr` probe instead code-switches ("Et so, my fellow Americans…", agreement 0.00) and loses, as it should. These VAD providers are available: **Silero VAD**(default) — ~885 KB, auto-downloaded via`--vad` . Industry-standard, well-tested.**FireRedVAD**— DFSMN-based, 2.4 MB, F1=97.57%. Pass`--vad -vm firered` to auto-download. Recommended.**MarbleNet**— NVIDIA 1D separable CNN, 439 KB, 6 languages (EN/DE/FR/ES/RU/ZH). Pass`--vad -vm marblenet` to auto-download. Smallest model. ()`cstr/marblenet-vad-GGUF` **Whisper-VAD-EncDec***(experimental)*— Whisper-base encoder + TransformerDecoder head, 22 MB Q4_K. Trained on Japanese ASMR; may not generalise well to all domains. Pass`--vad -vm whisper-vad` . Slower than others (~1s vs ~50ms). ()`cstr/whisper-vad-encdec-asmr-GGUF` Pass `--lid-backend off` to skip LID entirely. Audio LID (above) tags **what was spoken**; text LID tags **what was written**. Text LID runs on a transcript or any UTF-8 string and is useful for routing post-ASR pipelines (translation, punctuation, sub selection) without re-running an audio model. Three GGUF families, one binary — the dispatcher picks by `general.architecture` : | Backend | Labels | Size (F16) | License | HF repo | |---|---|---|---|---| CLD3 (Google compact language detector v3) | 109 ISO 639-1 | 440 KB | Apache-2.0 | `cstr/cld3-GGUF` | **GlotLID-V3**(cis-lmu fastText)`cstr/glotlid-GGUF` **LID-176**(Facebook fastText)`cstr/fasttext-lid176-GGUF` ¹ LID-176 is **CC-BY-NC-4.0** — non-commercial use only. CLD3 + GlotLID-V3 are Apache-2.0 with no such constraint. Pick CLD3 for the smallest, fastest path; GlotLID for maximum coverage (low-resource languages); LID-176 only if you need its specific 176-label space and accept its non-commercial terms. **Standalone CLI** — auto-routes by GGUF arch, with auto-download: ``` crispasr-lid -m auto --text "Bonjour le monde" # → cstr/cld3-GGUF (default, ~440 KB) crispasr-lid -m auto:glotlid --text "Bonjour le monde" -k 5 crispasr-lid -m auto:lid-fasttext176 --text "Hallo Welt" # Or pass an explicit path / canonical filename (looked up in the registry): crispasr-lid -m cld3-f16.gguf --text "你好世界" # zh 0.997816 echo "Привет мир" | crispasr-lid -m auto --quiet # ru 0.907322 ``` **Post-ASR pipeline** — `--lid-on-transcript` runs the same dispatcher on the assembled transcript (also accepts `auto[:variant]` ): ``` crispasr -m ggml-tiny.bin -f speech.wav --lid-on-transcript auto # (transcript on stdout) # lang=de conf=0.997123 backend=lid-cld3 ``` The dispatcher (`src/text_lid_dispatch.{h,cpp}` ) is a thin C ABI façade — one integer compare per call; per-stage diff harness is green at cos≥0.999 across 8 multilingual smoke samples. **Don't want to build?** Prebuilt binaries for Windows, macOS and Linux are on the [releases page](https://github.com/CrispStrobe/CrispASR/releases/latest) — see [Start here](#start-here) for which file to take. The rest of this section is for building from source. ``` git clone --recursive https://github.com/CrispStrobe/CrispASR cd CrispASR # already cloned without --recursive? initialize the bundled ggml submodule: # git submodule update --init --recursive cmake -B build -DCMAKE_BUILD_TYPE=Release cmake --build build -j$(nproc) ``` The `ggml/` submodule is required. If you cloned without `--recursive` , run `git submodule update --init --recursive` first — otherwise CMake stops with a message telling you to do exactly that. Produces `build/bin/crispasr` (main CLI), `build/bin/crispasr-quantize` , and `build/bin/crispasr-diff` . No Python, PyTorch, or pip required at runtime — just a C++17 compiler and CMake 3.14+. For GPU acceleration, add the matching ggml flag at configure time: ``` cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_CUDA=ON # NVIDIA cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_METAL=ON # Apple Silicon cmake -B build -DCMAKE_BUILD_TYPE=Release -DGGML_VULKAN=ON # cross-vendor ``` **See ** for the full guide: all GPU backends (CUDA / Metal / Vulkan / MUSA / SYCL), Windows convenience scripts, ffmpeg ingestion, optional BLAS, glibc notes, and the `docs/install.md` `scripts/dev-build.sh` wrapper.If a build runs but the binary exits with no output, see [ docs/troubleshooting.md](/CrispStrobe/CrispASR/blob/main/docs/troubleshooting.md). Deeper ASR examples below. If this is your first run, use [Start here](#start-here) instead. For TTS, the runnable guide is [docs/tts.md](/CrispStrobe/CrispASR/blob/main/docs/tts.md) ([Text-to-Speech](#text-to-speech-models) below is the model catalogue). ``` # Download a whisper model (same as upstream whisper.cpp) ./models/download-ggml-model.sh base.en ./build/bin/crispasr -m models/ggml-base.en.bin -f samples/jfk.wav # [00:00:00.000 --> 00:00:07.940] And so my fellow Americans ask not what your country can do for you # [00:00:07.940 --> 00:00:10.760] ask what you can do for your country. # Grab the quantized model (~467 MB) curl -L -o parakeet.gguf \ https://huggingface.co/cstr/parakeet-tdt-0.6b-v3-GGUF/resolve/main/parakeet-tdt-0.6b-v3-q4_k.gguf ./build/bin/crispasr -m parakeet.gguf -f samples/jfk.wav # Auto-detected backend 'parakeet' from GGUF metadata. # And so, my fellow Americans, ask not what your country can do for you, ask what you can do for your country. # Word-level timestamps (one line per word) ./build/bin/crispasr -m parakeet.gguf -f samples/jfk.wav -ml 1 # Transcription (source == target) ./build/bin/crispasr --backend canary -m canary-1b-v2-q5_0.gguf -f audio.de.wav -sl de -tl de # Translation (German speech → English text) ./build/bin/crispasr --backend canary -m canary-1b-v2-q5_0.gguf -f audio.de.wav -sl de -tl en # ...or use the familiar crispasr flag: ./build/bin/crispasr --backend canary -m canary-1b-v2-q5_0.gguf -f audio.de.wav -l de --translate # First run downloads ~2.5 GB to ~/.cache/crispasr/ via curl, then runs ./build/bin/crispasr --backend voxtral -m auto -f samples/jfk.wav # Subsequent runs use the cached file ./build/bin/crispasr --backend voxtral -m auto -f samples/jfk.wav -l en # 0.6B (default, ~500 MB) ./build/bin/crispasr --backend qwen3 -m auto -f audio.zh.wav # 1.7B (higher quality, ~1.3 GB) — supports both -hf and non-hf source models ./build/bin/crispasr --backend qwen3 -m qwen3-1.7b --auto-download -f audio.wav # Japanese anime/galgame fine-tune (~1.3 GB) ./build/bin/crispasr --backend qwen3 -m qwen3-ja-anime --auto-download -f anime.wav ``` **Long audio:** the default is safe 30 s chunking. `--chunk-seconds 0` decodes the whole file in ONE pass (matches the reference model verbatim on multi-minute clips, #218) — but the encoder's full attention is O(N²) in audio length, so keep single-pass clips under ~10 minutes on 16 GB machines. For long-form use prefer the plain `-q4_k` /`-q8_0` GGUFs over the `-imatrix` variants (see the model card). ``` ./build/bin/crispasr --backend glm-asr -m auto -f audio.wav # Long audio in one pass (up to 655 s — 30 s encoder windows, one LLM prompt, # same layout as the HF/zai reference; matches it verbatim on the #218 clip): ./build/bin/crispasr --backend glm-asr -m auto --chunk-seconds 0 -f long.wav ``` Note: in single-pass mode the model (like the reference) skips leading non-speech audio; the default 30 s-chunked mode transcribes more of such clips. Custom `--ask` / non-English `--language` instructions need a GGUF with baked BPE merges (re-published 2026-07; older GGUFs fall back to the default transcription prompt with a warning). ``` # Download the LM + audio tokenizer (the tokenizer is a separate model) huggingface-cli download cstr/mimo-asr-GGUF mimo-asr-q4_k.gguf \ --local-dir ~/.cache/crispasr huggingface-cli download cstr/mimo-tokenizer-GGUF mimo-tokenizer-q4_k.gguf \ --local-dir ~/.cache/crispasr # Transcribe (auto-discovers tokenizer if it sits next to the LM) ./build/bin/crispasr \ --backend mimo-asr \ -m ~/.cache/crispasr/mimo-asr-q4_k.gguf \ --codec-model ~/.cache/crispasr/mimo-tokenizer-q4_k.gguf \ -f samples/jfk.wav # Output: And so, my fellow Americans, ask not what your country can do # for you. Ask what you can do for your country. ``` The 4.5 GB Q4_K is the recommended quant; F16 (14.9 GB) needs ~16 GB RAM during inference. JFK matches the upstream Python `MimoAudio.asr_sft` reference verbatim; performance on M1+Metal is ~0.3× realtime (Q4_K dequant per step is the bottleneck — F16 + KV-reuse follow-ups are queued under PLAN #51a/b/c). ``` # English (Q4_K quantized, 212 MB — 6x smaller than F16) curl -L -o wav2vec2-en-q4k.gguf \ https://huggingface.co/cstr/wav2vec2-large-xlsr-53-english-GGUF/resolve/main/wav2vec2-xlsr-en-q4_k.gguf ./build/bin/crispasr -m wav2vec2-en-q4k.gguf -f samples/jfk.wav # and so my fellow americans ask not what your country can do for you ask what you can do for your country # German curl -L -o wav2vec2-de-q4k.gguf \ https://huggingface.co/cstr/wav2vec2-large-xlsr-53-german-GGUF/resolve/main/wav2vec2-xlsr-de-q4_k.gguf ./build/bin/crispasr -m wav2vec2-de-q4k.gguf -f audio.de.wav # Convert any HuggingFace Wav2Vec2ForCTC model: python models/convert-wav2vec2-to-gguf.py \ --model-dir jonatasgrosman/wav2vec2-large-xlsr-53-german \ --output wav2vec2-de.gguf --dtype f32 # Then optionally quantize: ./build/bin/crispasr-quantize wav2vec2-de.gguf wav2vec2-de-q4k.gguf q4_k ``` CrispASR has three feature areas that warrant their own docs pages: —[Streaming & live transcription](/CrispStrobe/CrispASR/blob/main/docs/streaming.md)`--stream` ,`--mic` ,`--live` , sliding-window chunking, per-token confidence.— Kokoro (multilingual, smallest), Qwen3-TTS (highest fidelity, voice cloning), VibeVoice (lowest-latency streaming), Orpheus (3 B Llama + SNAC), Chatterbox (flow-matching + HiFT vocoder, German via Kartoffelbox), IndexTTS, VoxCPM2, and CosyVoice3 (9 langs + 18 zh dialects; baked-voice bank + arbitrary-WAV cloning). Voice packs, language routing, and qwen3-tts environment switches. All TTS output is watermarked; post-embed verification warns if confidence is low. Use[Text-to-Speech (TTS)](/CrispStrobe/CrispASR/blob/main/docs/tts.md)`--detect-watermark file.wav` to check any WAV for AI watermarks.— persistent model, OpenAI-compatible[Server mode (HTTP API)](/CrispStrobe/CrispASR/blob/main/docs/server.md)`/v1/audio/transcriptions` (ASR) and`/v1/audio/speech` +`/v1/voices` (TTS, automatic on any loaded CAP_TTS backend), per-request voice + speed + instructions, CORS, long-form sentence chunking, API keys, Docker Compose, prebuilt CUDA images.— one transcription already uses multiple cores; the server accepts requests concurrently but serializes inference on one model by default;[Concurrency, parallelism & scaling](/CrispStrobe/CrispASR/blob/main/docs/concurrency.md)`--server-workers N` runs N model instances so pure-ASR requests run concurrently; and for bulk/throughput workloads, process-level fan-out (`xargs -P` / GNU`parallel` ) or N replicas behind a load balancer. Also covers what is*not*supported (batched multi-stream inference, PagedAttention) and why. Quickest taste of each: ``` # Streaming from microphone crispasr --mic -m model.gguf # TTS via auto-downloaded VibeVoice (~636 MB on first run) crispasr --backend vibevoice-tts -m auto --tts "Hello world" --tts-output hello.wav # CosyVoice3 on GPU; companions auto-download beside the LLM crispasr --backend cosyvoice3-tts -m auto --tts "Hello world" --tts-output cosy.wav # CosyVoice3 fast mode: 5 flow steps instead of the quality-default 10 COSYVOICE3_FLOW_STEPS=5 crispasr --backend cosyvoice3-tts -m auto \ --tts "Hello world" --tts-output cosy-fast.wav # Persistent HTTP server, OpenAI-compatible crispasr --server -m model.gguf --port 8080 curl -F "file=@audio.wav" http://localhost:8080/v1/audio/transcriptions # TTS over HTTP — load a TTS backend, hit /v1/audio/speech crispasr --server --backend qwen3-tts-customvoice -m auto --voice-dir ./voices --port 8080 curl http://localhost:8080/v1/audio/speech \ -H 'Content-Type: application/json' \ -d '{"input":"Hello world","voice":"vivian"}' -o out.wav ``` CosyVoice3 uses batched classifier-free guidance and request-sized KV caching by default. Baked voices load only the LLM, flow, HiFT, and voice bank; the larger S3 tokenizer and CAMPPlus companions load lazily when a `.wav` cloning voice is first requested. Common flags: ``` crispasr -m auto --backend parakeet -f audio.wav --vad -osrt --split-on-punct ``` | Flag | Meaning | |---|---| `-m FNAME` / `--backend NAME` | Model path (or `auto` ) and forced backend | `-f FNAME` | Input audio (repeatable; positional accepted) | `--vad` | Silero VAD chunking — strongly recommended for multi-minute audio | `-osrt` / `-ovtt` / `-otxt` / `-oj` / `-ojf` | Output formats (also `-ocsv` , `-olrc` ) | `-am FNAME` | CTC aligner GGUF for word-level timestamps on LLM backends | `--align-only` | Standalone forced alignment: text/`.srt` + audio → timestamped SRT/JSON (no ASR needed); `.srt` input keeps its cues and gets re-timed (`--align-granularity auto|word|segment` ) | `-tp F` / `-bs N` | Sampling temperature / beam search width | `-n N` / `--frequency-penalty F` | Generated-token cap / opt-in repeated-token penalty for supported autoregressive ASR backends | `-l auto` / `--detect-language` | LID pre-step for backends without native lang detect | `--hotwords "A,B,C"` | Contextual biasing — boost named terms during CTC/TDT decode or LLM prompt | `-ck N` | Fallback chunk size when VAD is off (default 30 s) | `--list-backends` | Print the capability matrix and exit | **See ** for the full reference: every flag, VAD details, CTC alignment workflow, output JSON layout, the auto-download registry, and supported audio formats. `docs/cli.md` **See** for Python / Rust / Dart / Go / Java / JavaScript / Ruby / mobile. `docs/bindings.md` CrispASR is structured as a stable C-ABI in `src/` (every algorithm: VAD, diarize, LID, alignment, cache, registry) consumed by all language wrappers, with thin presentation layers in `examples/cli/` . Per-model runtimes live in `src/{whisper,parakeet,canary,...}.cpp` , sharing primitives from `src/core/` (mel, ffn, attention, GGUF loader, FastConformer / Conformer / Granite-LLM blocks, etc.). — full layered layout, file-by-file tour of`docs/architecture.md` `src/` and`examples/cli/` , per-backend internals table, regression discipline.— adding a new backend in five files, clang-format-18 setup, the`docs/contributing.md` `crispasr-diff` PyTorch-ground-truth workflow, and the TTS audio-cosine-vs-reference regression target.—`docs/regression-matrix.md` `tools/test-all-backends.py` capability tiers, cache modes (`keep` /`ephemeral` ),`--skip-missing` for CI. **Shared libraries** (cross-repo with CrispEmbed): `crisp_audio/` — Whisper-shape audio encoder (Conv-stem + Transformer)`crisp_punc/` — punctuation restoration (FireRedPunc + PCS)`crisp_lid/` — text-based language identification (fastText + CLD3)`crisp_truecase/` — truecasing (statistical + CRF + BiLSTM) Both are self-contained static libraries with CMakeLists.txt. CrispEmbed links them via `add_subdirectory(../CrispASR/crisp_*/)` ; CrispASR uses them directly. If the shared dir is absent, both repos fall back to local copies of the source files. For benchmarks see [ PERFORMANCE.md](/CrispStrobe/CrispASR/blob/main/PERFORMANCE.md); for the session-by-session port log and the bug-class lessons, see [.](/CrispStrobe/CrispASR/blob/main/LEARNINGS.md) `LEARNINGS.md` `build/bin/crispasr-quantize` is a single, model-agnostic GGUF re-quantization tool that works across all supported model families (Whisper, Parakeet, Canary, Cohere, Voxtral, Qwen3, Granite, Wav2Vec2, MiMo-ASR, GLM-ASR, Moonshine, VibeVoice, Kokoro, Qwen3-TTS, …): ``` ./build/bin/crispasr-quantize input.gguf output.gguf q4_k ``` **See ** for the full guide: supported quant types, K-quant alignment fallback, recommended quant per backend, and worked examples for each architecture. `docs/quantize.md` All backends use `ggml_backend_init_best()` which automatically picks the highest-priority compiled backend: CUDA > Metal > Vulkan > CPU. To force a specific backend: ``` # Force Vulkan even when CUDA is available crispasr --gpu-backend vulkan -m model.gguf -f audio.wav # Pin a specific GPU (useful on Vulkan systems with iGPU + dGPU) crispasr --gpu-backend vulkan -dev 1 -m model.gguf -f audio.wav # Force CPU (useful for benchmarking) crispasr -ng -m model.gguf -f audio.wav # CUDA unified memory (swap to RAM when VRAM exhausted) GGML_CUDA_ENABLE_UNIFIED_MEMORY=1 crispasr -m model.gguf -f audio.wav ``` Build flags: `-DGGML_CUDA=ON` , `-DGGML_METAL=ON` , `-DGGML_VULKAN=ON` . Notes: `--gpu-backend vulkan` selects the Vulkan backend, but it does not choose which physical GPU to use. Use`-dev N` to select the Vulkan device index.- On some Windows laptops, Vulkan device `0` is the Intel iGPU and the NVIDIA GPU is`1` . If Vulkan looks unexpectedly slow, rerun with`-dev 1` . - The Windows convenience script `build-vulkan.bat` creates a separate Vulkan-capable binary at`build-vulkan\bin\crispasr.exe` . For most backends, `-v` / `--verbose` surfaces per-stage timings and device picks. For headless / library use (where the CLI flag isn't plumbed through), set `CRISPASR_VERBOSE=1` instead. ``` # Per-stage timing breakdown (mel / encoder / prefill / decode): crispasr -v --backend gemma4-e2b -m model.gguf -f audio.wav # gemma4_e2b: mel 128x1099 (17.2 ms) # gemma4_e2b: encoder done: 1536x275 (719.0 ms) # gemma4_e2b: prefill done, first_token=3133 (1464.0 ms) # gemma4_e2b: decoded 25 tokens (7748.3 ms total) # crispasr: transcribed 11.0s audio in 7.75s (1.4x realtime) # Hugging Face access for gated models (Voxtral, Gemma4-E2B, …): HF_TOKEN=hf_xxx crispasr -m auto --backend gemma4-e2b -f audio.wav ``` The server has its own auth env: `CRISPASR_API_KEYS` (see [Server mode](/CrispStrobe/CrispASR/blob/main/docs/server.md)). **Per-backend debug / bench / dump-dir env vars (developer)** These are useful when porting a new backend or chasing a regression. The `*_BENCH=1` toggles emit per-stage timings even without `-v` ; the `*_DEBUG=1` toggles emit per-step diagnostic prints; the `*_DUMP_DIR=` paths write per-stage F32 tensors for diff-testing against a PyTorch reference (see [Debug a new backend against PyTorch ground truth](/CrispStrobe/CrispASR/blob/main/docs/contributing.md#debug-a-new-backend-against-pytorch-ground-truth)). | Env var | Purpose | |---|---| `CRISPASR_VERBOSE=1` | Forces verbose mode for any backend (parallel to the `-v` flag). | `CRISPASR_DUMP_DIR=path/` | Generic per-stage F32 tensor dump for the `crispasr-diff` harness. | `GEMMA4_E2B_BENCH=1` | Per-stage timings for the Gemma-4-E2B backend. | `COHERE_BENCH=1` / `COHERE_DEBUG=1` | Cohere transcribe per-stage timings / per-step diagnostics. | `COHERE_PROF=1` | Cohere graph-level profiling (per-op timings). | `COHERE_THREADS=N` | Override thread count for the Cohere backend. | `COHERE_DEVICE=cpu|cuda|metal|vulkan` | Force the Cohere backend onto a specific device. | `COHERE_DUMP_ATTN=path/` | Dump attention activations for Cohere (used by the diff harness). | `FIRERED_BENCH=1` | Per-stage timings for the FireRedASR backend. | `FIREREDPUNC_DEBUG=1` | Per-step diagnostics for the FireRed punctuation post-step. | `MOONSHINE_STREAMING_BENCH=1` | Per-stage timings for moonshine-streaming. | `OMNIASR_BENCH=1` / `OMNIASR_DEBUG=1` / `OMNIASR_DUMP_DIR=` | OmniASR per-stage timings, diagnostics, and stage dumps. | `PARAKEET_DEBUG=1` | Parakeet TDT per-step diagnostics (joint network, blank-id sanity). | `QWEN3_TTS_BENCH=1` / `QWEN3_TTS_DEBUG=1` / `QWEN3_TTS_DUMP_DIR=` | Qwen3-TTS per-stage timings, diagnostics, and stage dumps. | `VIBEVOICE_BENCH=1` / `VIBEVOICE_DEBUG=1` / `VIBEVOICE_DUMP_DIR=` | VibeVoice ASR per-stage timings, diagnostics, and stage dumps. | `VIBEVOICE_REF_FEATURES=path` | Replace the live encoder with a saved feature tensor (regression harness). | `VIBEVOICE_TTS_DUMP=path/` | VibeVoice TTS per-stage dumps (token IDs, base/TTS hidden, neg condition, frame-0 noise/v_cfg/latent/acoustic_embed) for the diff harness. | `VIBEVOICE_TTS_DUMP_PERFRAME=1` | Per-frame VibeVoice TTS dumps written as `perframe__f.bin` . Pair with `VIBEVOICE_TTS_DUMP=path/` and `VIBEVOICE_TTS_NOISE=path` for stage-by-stage AR diff against `tools/run_official_vibevoice.py` . | `VIBEVOICE_TTS_TRACE=1` | Extra one-line traces (negative-condition prefill rms, scaling/bias factors loaded). Same effect as library verbosity ≥ 2; there is no CLI flag for it (`-v` caps verbosity at 1). | `VIBEVOICE_VOICE_AUDIO=path.wav` | Reference voice WAV for 1.5B-base TTS without a `.gguf` voice cache. | `VIBEVOICE_TTS_NOISE=path` | Override the per-frame Gaussian init noise. Flat little-endian float32 `[N_frames, vae_dim]` — typically the `noise.bin` written by `tools/run_official_vibevoice.py` . | `VIBEVOICE_VAE_BACKEND=cpu|metal|cuda|vulkan` | Pin the VAE decoder onto a specific backend. | `WAV2VEC2_BENCH=1` / `WAV2VEC2_VERBOSE=1` / `WAV2VEC2_DUMP_DIR=` | wav2vec2 per-stage timings, verbose graph traces, and stage dumps. | `CRISPASR_VOXTRAL4B_STREAM_TIMING=1` | Per-stage timings for the voxtral4b streaming path (encoder drain / prefill / first-text-token / decode-step p50/p95). | `CRISPASR_VOXTRAL4B_STREAM_CHUNK_MS=N` | Override the internal encoder chunk size (default 240 ms). Must be a multiple of 80 ms. Larger = faster feed (kernel-launch amortisation), longer live-caption latency floor. | `CRISPASR_VOXTRAL4B_STREAM_BATCH_ENCODER=1` | Regression-debug: ignore the streaming encoder's audio_embeds and re-run the whole batch encoder at flush. | `CRISPASR_VOXTRAL4B_STREAM_DEBUG=1` / `CRISPASR_VOXTRAL4B_STREAM_DIFF=1` | Per-step decode prints / side-by-side encoder cosine vs the batch encoder. | `CRISPASR_VOXTRAL4B_STREAM_LIVE=1` | Live-captions decode-during-feed (PLAN #7 phase 3). `get_text()` polled during feed returns progressive transcript. Default OFF (PTT semantics). Wrappers: Python `Session.stream_open(live=True)` , Rust `stream_open_ex(.., live: true)` . | `CRISPASR_VOXTRAL4B_STREAM_DECODER_THREAD=1` | Decoder worker thread (PLAN #7 phase 4, implies live mode). Lets `feed()` return between encoder chunks without waiting for the decode loop — useful for mic-driven workloads. On M1 the Metal queue serializes encoder and decoder so total wall-clock is unchanged; faster GPUs with kernel-level parallelism see real overlap. | `CRISPASR_VOXTRAL4B_FUSED_QKV=0` | Opt out of the runtime fused-QKV LLM path (default-on, ~7-8 % decode speedup on M1 Q4_K, ~500 MB extra memory). | `CRISPASR_QWEN3_ASR_FUSED_QKV=0` | Opt out of the runtime fused-QKV LLM path for qwen3-asr (default-on; works on F16/F32/Q4_K/Q8_0/...). | `CRISPASR_VOXTRAL_FUSED_QKV=1` | Opt in to the runtime fused-QKV LLM path for voxtral 3B. Off by default (no measurable speedup on JFK-shape decodes; useful for long-form workloads where decode dominates). | `QWEN3_TTS_FUSED_QKV=1` | Opt in to the runtime fused-QKV talker path. | `GRANITE_DISABLE_ENCODER_GRAPH=1` | Force the granite-speech / -plus / -nar encoder back to the per-layer CPU loop (slower but kept around for debugging). The single ggml-graph encoder with per-layer Shaw RPE is the default and is bit-near-identical to the CPU loop while being ~2× faster end-to-end across all three variants. | `CRISPASR_NO_REL_POS=1` | Ablate the relative-position bias in the Gemma-4 audio encoder (development only). | `ECAPA_REF_FBANK=path` | Reference filterbank tensor for the ECAPA-TDNN LID model (regression harness). | `CRISPASR_SHERPA_LID_BIN=path` | Override the auto-detected sherpa-onnx LID binary. | `CRISPASR_ARG_DEVICE=N` | Default GPU device index when `-dev` isn't passed. | `GGML_CUDA_ENABLE_UNIFIED_MEMORY=1` | Let CUDA swap to RAM when VRAM is exhausted. | `GGML_VK_VISIBLE_DEVICES` / `CUDA_VISIBLE_DEVICES` | Standard ggml/CUDA device-visibility filters. | `HF_TOKEN` and `HUGGING_FACE_HUB_TOKEN` are both honoured for gated-model downloads (in that order). — the original ggml inference engine and Whisper runtime this fork is built on[whisper.cpp](https://github.com/ggml-org/whisper.cpp)— the tensor library everything runs on[ggml](https://github.com/ggml-org/ggml)**NVIDIA NeMo**— parakeet-tdt-{0.6b-v2,0.6b-v3,1.1b}, parakeet-tdt_ctc-{110m,1.1b,0.6b-ja}, parakeet-ctc-{0.6b,1.1b}, canary-1b-v2, canary-ctc aligner, and the FastConformer-CTC family (stt_en_fastconformer_ctc_{large,xlarge,xxlarge} plus CTC branches of the stt_*_fastconformer_hybrid_large[_pc] fleet: en-pc, de, es, fr, it, nl, pl, ru, ua, hr, be, ar, fa, ka, hy, uz, kk-ru — all usable both as ASR backends and as compact ~82 MB`-am` forced aligners)**Cohere**— cohere-transcribe-03-2026** Qwen team (Alibaba)**— Qwen3-ASR-0.6B, Qwen3-ASR-1.7B, Qwen3-ForcedAligner-0.6B** Mistral AI**— Voxtral Mini 3B and 4B Realtime** IBM Granite team**— Granite Speech 3.2-8b, 3.3-2b, 3.3-8b, 4.0-1b** Meta / wav2vec2**— wav2vec2 CTC models (XLSR-53 English, German, multilingual via any Wav2Vec2ForCTC checkpoint)— optional diarization via subprocess (ONNX models)[sherpa-onnx](https://github.com/k2-fsa/sherpa-onnx)— VAD (native GGUF) and language identification (native GGUF, 95 languages)[Silero](https://github.com/snakers4/silero-vad)— speaker diarization segmentation (native GGUF port)[pyannote](https://github.com/pyannote/pyannote-audio)+[miniaudio](https://miniaud.io/)+[stb_vorbis](https://github.com/nothings/stb)— embedded/linked audio decoders (WAV/MP3/FLAC/AIFF/OGG/Opus, no ffmpeg; AAC/M4A/ALAC via Apple AudioToolbox)[libopus/opusfile](https://opus-codec.org/)(MIT) — in-tree clean-room codec suite: MP3/AAC-LC/Ogg-Opus output encoders (TTS[glint](https://github.com/CrispStrobe/glint)`.mp3` /`.aac` /`.opus` ) and cross-platform ADTS AAC-LC + Ogg Opus input decoders (no libopus needed)(Anthropic) — significant portions of the crispasr integration layer, all model converters, and the FastConformer/attention/mel/FFN/BPE core helpers were co-authored with Claude[Claude Code](https://claude.ai/claude-code) Same as upstream whisper.cpp: **MIT**. Per-model weights are covered by their respective HuggingFace model licenses (see [Supported backends](#supported-backends)). The `crispasr` binary itself links model runtimes that are mostly permissively licensed (MIT / Apache-2.0 / CC-BY-4.0 for weights).