FreePublished on GitHub

funasr

Router for FunASR speech workflows: local transcription, subtitles, services, vLLM, normalization, and training/export.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill

See how to install

Files included (40)

SKILL.md4.3 KB
references/api-reference.md4.3 KB
references/cli-reference.md3.1 KB
references/data-formats.md2.4 KB
references/model-overview.md3.2 KB
references/repo-provenance.md861 B
references/repo-routing-metadata.json509 B
references/troubleshooting.md2.8 KB
references/workflows.md3.0 KB
scripts/check_funasr_env.py10.9 KB
sub-skills/llm-asr-and-vllm/references/troubleshooting.md9.1 KB
sub-skills/llm-asr-and-vllm/references/vllm-and-models.md10.2 KB
sub-skills/llm-asr-and-vllm/scripts/check_vllm_ready.py22.4 KB
sub-skills/llm-asr-and-vllm/SKILL.md4.8 KB
sub-skills/python-asr-pipelines/references/api-reference.md6.2 KB
sub-skills/python-asr-pipelines/references/cli-reference.md2.9 KB
sub-skills/python-asr-pipelines/references/troubleshooting.md3.2 KB
sub-skills/python-asr-pipelines/references/workflows.md4.1 KB
sub-skills/python-asr-pipelines/scripts/batch_transcribe.py12.6 KB
sub-skills/python-asr-pipelines/scripts/generate_subtitles.py10.7 KB
sub-skills/python-asr-pipelines/SKILL.md2.2 KB
sub-skills/serving-and-runtime/references/mcp-and-client-integrations.md3.3 KB
sub-skills/serving-and-runtime/references/openai-api.md3.9 KB
sub-skills/serving-and-runtime/references/realtime-websocket.md4.7 KB
sub-skills/serving-and-runtime/references/runtime-sdk-and-edge.md5.2 KB
sub-skills/serving-and-runtime/references/troubleshooting.md4.9 KB
sub-skills/serving-and-runtime/scripts/funasr_mcp_server.py8.6 KB
sub-skills/serving-and-runtime/scripts/openai_api_smoke_test.py4.7 KB
sub-skills/serving-and-runtime/SKILL.md3.4 KB
sub-skills/text-normalization/references/troubleshooting.md4.3 KB
sub-skills/text-normalization/references/workflows.md5.1 KB
sub-skills/text-normalization/scripts/post_process_punct.py10.6 KB
sub-skills/text-normalization/SKILL.md3.2 KB
sub-skills/training-data-and-export/references/data-formats.md5.4 KB
sub-skills/training-data-and-export/references/export-and-onnx.md4.6 KB
sub-skills/training-data-and-export/references/troubleshooting.md6.8 KB
sub-skills/training-data-and-export/references/workflows.md6.2 KB
sub-skills/training-data-and-export/scripts/make_jsonl_from_scp.py8.0 KB
sub-skills/training-data-and-export/scripts/validate_manifest.py11.2 KB
sub-skills/training-data-and-export/SKILL.md3.1 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

FunASR

Use this root skill to route FunASR requests. Start here when the user asks for speech transcription, subtitles, model choice, service deployment, LLM-ASR routing, punctuation cleanup, manifests, or training/export guidance.

Start here

Package provenance and staleness check: references/repo-provenance.md
Router metadata for repo-skills-router: references/repo-routing-metadata.json
Package commands and entry points: references/cli-reference.md
Public API shapes and input/output forms: references/api-reference.md
Model-family choice guide: references/model-overview.md
Cross-skill workflow map: references/workflows.md
Shared input/output schemas: references/data-formats.md
Cross-cutting failure recovery: references/troubleshooting.md
Safe environment smoke helper: scripts/check_funasr_env.py

Route map

Need	Go to
Plain transcription, batch ASR, subtitles, audio decoding, hotwords, timestamps, or common non-LLM model choice	sub-skills/python-asr-pipelines/SKILL.md
Punctuation cleanup, inverse text normalization, or text normalization	sub-skills/text-normalization/SKILL.md
OpenAI-compatible HTTP API, realtime WebSocket, MCP, browser/client integration, or edge/runtime guidance	sub-skills/serving-and-runtime/SKILL.md
Fun-ASR-Nano, Fun-ASR-MLT-Nano, GLM-ASR-Nano, Qwen3-ASR, or `AutoModelVLLM`	sub-skills/llm-asr-and-vllm/SKILL.md
Training data prep, manifests, distributed config, export, or local inference after export	sub-skills/training-data-and-export/SKILL.md

What this root skill owns

The first decision about which FunASR workflow to use.
Quick package inspection and safe help/version checks.
The common vocabulary for ASR, VAD, punctuation, speaker labels, hotwords, subtitles, service deployment, and manifests.
Cross-cutting troubleshooting that applies before a user narrows to a sub-skill.

How to start safely

Run python scripts/check_funasr_env.py --check-cli --check-torch to confirm the package and entry points are visible in the current environment.
Use import funasr as the minimal package import check.
Remember that import funasr may work before PyTorch is installed, but from funasr import AutoModel needs a compatible torch build.
If the request names a model family, use references/model-overview.md before choosing a route.

Where to read next

Transcription and subtitles: python-asr-pipelines
Optional punctuation cleanup or ITN/TN: text-normalization
HTTP/WebSocket/MCP/runtime deployment: serving-and-runtime
Nano/GLM/Qwen3/vLLM backend choice: llm-asr-and-vllm
Training and export: training-data-and-export

Do not do here

Do not bury service, model-family, normalization, or training guidance inside this router.
Do not tell users to open the original source checkout from runtime instructions.
Do not claim GPU, vLLM, or Pynini support from the root skill alone; route to the owning sub-skill and its references.

Quick model instincts

CPU-friendly first try: SenseVoiceSmall
Mandarin production and subtitles: paraformer-zh
GPU-heavy LLM-ASR / batch throughput: Fun-ASR-Nano or GLM-ASR-Nano through llm-asr-and-vllm
Server or client integration: serving-and-runtime

Cross-cutting checks

Use references/troubleshooting.md for install/import, hub, device, and download issues.
Use the sub-skill troubleshooting pages for workflow-specific failures.
Keep review-only outputs in skills/tests/funasr/, not inside the runtime skill tree.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

More skills from this repository

3d-resnets-pytorch

Free

Routes 3D ResNets PyTorch video action-recognition workflows across training, inference, and data preparation.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

3ddfa

Free

Guide 3DDFA Python inference, geometry rendering, training/evaluation, and optional C++ ONNX workflows for 3D dense face alignment.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

3ddfa-v2

Free

Routes 3DDFA_V2 face-alignment setup, still-image demos, video tracking, and ONNX benchmarking workflows.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

ab3dmot

Free

Operate AB3DMOT 3D multi-object tracking workflows for KITTI and nuScenes data, tracking, evaluation, and visualization.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

accelerate

Free

Use Hugging Face Accelerate for PyTorch training-loop migration, distributed launch/configuration, DeepSpeed/FSDP/TPU backend setup, big-model inference/offload, checkpointing, tracking, and troubleshooting.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

acme

Free

Route Acme reinforcement-learning framework tasks across core loops, replay/data, JAX agents, and TensorFlow/Sonnet agents.

The English summary is being prepared. Showing the original description.

VectorSpaceLab/AREX-Skill339Updated Sep 3, 2026

See all skills by VectorSpaceLab

Report a problem with this skill

Install

Install with a command

npx skills add https://github.com/VectorSpaceLab/AREX-Skill --skill funasr

Paste it into your terminal and run it (requires Node.js). You can choose which agents to install it for, such as Claude Code or Codex.

Ask your agent

Please install the following Agent Skill.
https://github.com/VectorSpaceLab/AREX-Skill/tree/main/skills/repositories/repo-skills/funasr

First review SKILL.md and the bundled files and confirm there is nothing suspicious (such as sending data externally or reading credentials). Then place it in ~/.claude/skills/funasr/ for Claude Code, or ~/.agents/skills/funasr/ for Codex.

Send this text to Claude Code or Codex as is. It checks the contents before installing.

View source on GitHub

Repository: VectorSpaceLab/AREX-Skill
Stars: 339
Forks: 26
Category: Productivity
License: Apache-2.0
Original language: English
Last updated: Sep 3, 2026

A skill is a set of instructions and scripts for an agent. cccskills does not guarantee that skills are safe. Review the contents before you install.