Menu
Local Qwen3-ASR speech-to-text provider for Ambient push-to-talk.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Safely observe the Ambient Blockchain, plan Tool Oracle and x402 interactions, prepare Ambient program deployment work, and run local live-test evidence gates.
Create reusable TinyStyler style profiles from user text examples and apply TinyStyler-style text transfer through Ambient CLI.
Generate raster images through hosted APIs such as OpenAI, Google Nano Banana Pro, fal/FLUX, Replicate, Stability AI, and Ideogram.
Inspect a workspace and summarize notable files for fixture-plugin tests.
Review workspace files for the Pi package fixture.
Local faster-whisper STT provider for Ambient speech input validation.
| name | ambient-qwen3-asr |
| description | Local Qwen3-ASR speech-to-text provider for Ambient push-to-talk. |
Use this package only through Ambient's STT provider path. It implements:
qwen3_asr_transcribe --audio <wav> --language <language> --output-json <path>
The command expects Ambient to provide a normalized 16 kHz mono WAV. It writes concise JSON with text, language, durationMs, providerId, and artifact metadata. It must not print raw audio, secrets, or long diagnostic payloads to stdout.
Runtime configuration:
AMBIENT_QWEN3_ASR_BINARY: optional path or command name for llama-mtmd-cli.assets/qwen3-asr-assets.json, downloaded into an Ambient cache on first use, verified by exact byte size and SHA-256, then passed to llama.cpp with -m and --mmproj.AMBIENT_QWEN3_ASR_ASSET_DIR: optional cache root for those verified manifest assets.AMBIENT_QWEN3_ASR_MODEL: optional custom llama.cpp -hf model ref for explicit experiments. Setting it bypasses the default verified manifest asset path.AMBIENT_QWEN3_ASR_THREADS: optional thread count. Defaults to a conservative local value.AMBIENT_QWEN3_ASR_FAKE_TRANSCRIPT: deterministic test hook. Do not set this for product use.Product constraints: