Skip to content

Open-source speech recognition toolkit for training, inference, streaming ASR, VAD, punctuation, speaker diarization pipelines, and OpenAI-compatible/MCP serving.

readme: improve conversion - human-readable output + Why FunASR comparison table

游雁b3e7a57b8de6
diff --git a/README_new_proposal.md b/README_new_proposal.md
index 5740180aa..af94de897 100644
--- a/README_new_proposal.md
+++ b/README_new_proposal.md
@@ -33,24 +33,36 @@ pip install funasr
```python
from funasr import AutoModel
-# One line: ASR + VAD + punctuation + speaker diarization
-model = AutoModel(
- model="iic/SenseVoiceSmall",
- vad_model="fsmn-vad",
- spk_model="cam++",
- device="cuda",
-)
+model = AutoModel(model="iic/SenseVoiceSmall", vad_model="fsmn-vad", spk_model="cam++", device="cuda")
result = model.generate(input="meeting.wav")
+```
-# result[0]["sentence_info"]:
-# [{"text": "Let's discuss the Q3 plan.", "start": 420, "end": 3800, "spk": 0},
-# {"text": "Sounds good. I have three points.", "start": 4200, "end": 7100, "spk": 1}]
+**Output** — structured text with speaker labels, timestamps, and punctuation:
+```
+[00:00.4 → 00:03.8] Speaker 0: Let's discuss the Q3 plan.
+[00:04.2 → 00:07.1] Speaker 1: Sounds good. I have three points.
+[00:07.5 → 00:12.3] Speaker 0: Go ahead. We have 30 minutes.
```
+That's it. **One model, one call** — VAD segmentation, speech recognition, punctuation, speaker diarization all happen automatically.
+
> **Deploy as API server:** `funasr-server --device cuda` → OpenAI-compatible endpoint at localhost:8000
>
> **Use with AI agents:** [MCP Server](examples/mcp_server/) for Claude/Cursor · [OpenAI API](examples/openai_api/) for LangChain/Dify/AutoGen
+### Why FunASR?
+
+| | FunASR | Whisper | Cloud APIs |
+|---|---|---|---|
+| Speed | **170x realtime** | 13x realtime | ~1x realtime |
+| Speaker ID | ✅ Built-in | ❌ Needs pyannote | ✅ Extra cost |
+| Emotion | ✅ Happy/Sad/Angry | ❌ | ❌ |
+| Languages | 50+ | 57 | Varies |
+| Streaming | ✅ WebSocket | ❌ | ✅ |
+| Self-hosted | ✅ MIT license | ✅ MIT license | ❌ Cloud only |
+| Cost | Free | Free | $0.006/min+ |
+| CPU viable | ✅ 17x realtime | ❌ Too slow | N/A |
+
---
<a name="benchmark"></a>