OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Audio & Voice AI

Yapsnap

Yapsnap is an open-source Python command-line tool for transcribing local audio or video files and supported video URLs into UTF-8 plaintext using CPU-based speech recognition. It uses sherpa-onnx with a streaming Zipformer transducer (Kroko); yt-dlp fetches URL media and ffmpeg decodes it. The tool supports automatic language detection, optional sentence timestamps and speaker diarization, and caches its models for

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Yapsnap — OpenViST project profile in Audio & Voice AI.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Yapsnap when you can work from a command line and want to transcribe local media or a URL handled by yt-dlp to a text file, with speech recognition performed on your CPU.

Who it is for

  • People who need text transcripts from local recordings or supported online videos.
  • Developers and other terminal users who want CPU-based transcription on macOS, Linux or Windows.
  • Users who want optional sentence timestamps or speaker labels in transcripts.

Documented uses

  • Turn an audio or video file into a plaintext transcript.
  • Extract and transcribe audio from a video URL supported by yt-dlp.
  • Save a transcript to a chosen file path.
  • Add sentence-level timestamps or speaker labels to a transcript.
  • Transcribe in a supported language using automatic detection or a manually selected model.
Before you choose

Requirements & limits

Yapsnap is documented as a public GitHub project under the Apache-2.0 license. The README describes Python 3.9 or later and support for macOS, Linux and Windows; ffmpeg is a prerequisite. Installation is documented through PyPI or from a repository clone.

  • The primary workflow is a terminal command, not a hosted web application.
  • Speech recognition is CPU-based; no GPU is required.
  • The first run downloads model files, approximately 20 MB for language identification and 80 MB per language model, which are cached for later runs.
  • Automatic language detection is documented for German, English, Spanish, French, Italian, Hebrew, Dutch, Portuguese, Swedish and Turkish.
  • Video URL support depends on yt-dlp extractors, which may change as source websites change.

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

Install Yapsnap and transcribe a local file

Install the command-line tool and create a plaintext transcript from a local media file.

Before you start

  • Python 3.9 or later.
  • ffmpeg installed and available on PATH.
  • A local media file that ffmpeg can decode.
  • Internet access for the initial model downloads.

Steps

  1. Install ffmpeg for your operating system and ensure it is on PATH.

  2. Install Yapsnap with the command below.

    Documented command
    pip install yapsnap

  3. Run yapsnap path/to/audio.mp3 or another supported file path.

  4. Find the UTF-8 transcript in ./transcripts/, or specify a destination with -o. Add --timestamps or --diarize if those transcript annotations are wanted.

Expected outcome

A plaintext transcript file, optionally with sentence timestamps or speaker labels.

Guide limitations

  • The first run downloads a language-identification model and a speech-recognition model; later runs can work offline.
  • Diarization labels are assigned in order of appearance and do not identify speakers across different files.
  • Supported input must be decodable by ffmpeg.
Documentation 1 ↗
Source-based practical guide

Transcribe a video URL

Fetch audio from a supported video URL and save its transcript as plaintext.

Before you start

  • Python 3.9 or later.
  • ffmpeg installed and available on PATH.
  • Yapsnap installed with pip install yapsnap.
  • A URL handled by yt-dlp and internet access for fetching media and initial model downloads.

Steps

  1. Run yapsnap "VIDEO_URL" with a URL supported by yt-dlp.

  2. Use -o PATH to choose a transcript destination, or accept the default transcript file location.

  3. For later transcription runs, use the cached speech-recognition model; optionally add --timestamps or --diarize when needed.

Expected outcome

A UTF-8 plaintext transcript saved as a text file.

Guide limitations

  • Website extractor changes can cause a URL to stop working; the README suggests updating yt-dlp if a URL begins to fail.
  • Models download on first use; later runs can be offline, but URL fetching itself requires network access.
  • URL access and media availability depend on the source website.
Documentation 1 ↗
Evidence and transparency

Sources & review

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