OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Image & Video AI

Funclip

FunClip is an open-source Python application for transcribing video with FunASR and creating clips and subtitles from recognized speech. Its Gradio interface supports manual selection of transcript segments or speakers and LLM-assisted clip selection, with local deployment or browser access to a server-hosted instance.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Funclip — OpenViST project profile in Image & Video AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use FunClip when a video workflow benefits from speech recognition first, followed by transcript- or speaker-based clip selection and subtitle generation. It is suited to local use or browser access to a server-hosted Gradio interface.

Who it is for

  • Video creators who want to select clips from recognized speech rather than locate every segment manually.
  • Teams seeking a locally deployed transcription and subtitle workflow.
  • Developers interested in adapting a Python and Gradio application built around FunASR speech recognition.

Documented uses

  • Transcribe video and obtain timestamped recognition results.
  • Select text segments or recognized speakers and produce corresponding video clips.
  • Generate full-video and target-segment SRT subtitles.
  • Use hotwords to guide speech recognition toward specified names or terms.
Before you choose

Requirements & limits

The repository is public and lists an MIT license. FunClip is described as locally deployable and as deployable on a server for browser access. Its README also links to hosted ModelScope and Hugging Face quick experiences.

  • Installation instructions specify Python 3.12 and a matching PyTorch/torchaudio pair for the platform.
  • The project says to retain its Transformers 4.x requirements rather than use Nano's separate Transformers 5.x quickstart.
  • Current model and subtitle compatibility paths require FunASR 1.4.9 or newer.
  • Model weights download separately at startup and are not included in the source archives.
  • The README documents Paraformer-based recognition, hotword customization through SeACo-Paraformer, speaker recognition through CAM++, and LLM-assisted clipping.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Prepare a Python environment for FunClip

Clone the repository, install its requirements in a separate Python 3.12 environment, and check the installation and launcher help.

Before you start

  • Python 3.12
  • A matching PyTorch/torchaudio pair for the platform
  • Git
  • A shell environment that supports Python virtual environments

Steps

  1. Clone the FunClip repository and change into its directory.

  2. Create and activate a Python 3.12 virtual environment.

  3. Install the repository requirements and run the package consistency check.

  4. Run the launcher with --help to view its help output.

Expected outcome

A prepared FunClip Python environment and the launcher's help output; this guide does not cover starting or using the Gradio service.

Guide limitations

  • Keep the environment on the repository's Transformers 4.x requirements; Nano's separate Transformers 5.x quickstart is not an upgrade instruction for FunClip.
  • Current model and subtitle compatibility paths require funasr>=1.4.9.
  • Model weights are downloaded separately when FunClip starts.
  • The documented commands do not establish that the application has been successfully run on a particular machine.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
7 October 2026
Original website title
modelscope/FunClip · Official homepage title · retained in its source language
How we review sourcesSuggest a correction