Feature: Auto-generate subtitles with FunASR/SenseVoice speech recognition

[LauraGPT]

Hi! Great open-source video editor!

Would you consider adding automatic subtitle generation powered by FunASR or SenseVoice?

Why?

Many video editors (DaVinci Resolve, Premiere Pro) have added AI subtitle generation. FunASR/SenseVoice could bring this to OpenShot:

• Ultra-fast — SenseVoice-Small: ~70ms for 10s audio
• Built-in punctuation — Natural sentence boundaries for subtitles
• Character-level timestamps — Precise subtitle timing
• 50+ languages — Wide coverage
• Fully offline — No cloud API needed
• Apache 2.0 — Compatible with OpenShot's open-source philosophy

Simple Python integration

from funasr import AutoModel

model = AutoModel(model="paraformer-zh", vad_model="fsmn-vad", pnc_model="ct-punc")
result = model.generate(input="video_audio.wav")
# Returns: text + timestamps → convert to SRT

References

• FunASR: https://github.com/modelscope/FunASR (16K+ stars)
• SenseVoice: https://github.com/FunAudioLLM/SenseVoice (8K+ stars)

[Colorjet3]

Hello @LauraGPT Thank you for your enhancement submission. I will assign this to the lead developer for his review. He will then Approve/Reject the request. If Approved, it will be added to the "Enhancement" queue for a future implementation.

[LauraGPT]

Thank you for routing it for review. That sounds good to me.

If the lead developer decides this fits the enhancement queue, I’m happy to help clarify the FunASR/SenseVoice integration options or test an early implementation with subtitle-generation samples.