Categories
Self-Hosted Productivity Note Taking Whisper AI Summarization Speech To Text Podcast Tools Video TranscriptionStars
Forks
Watchers
Developer links
AI Video Transcriber
A two-hour podcast becomes a page of takeaways with AI Video Transcriber, a self-hosted web app that turns links and uploads into clean transcripts and short summaries. Paste a URL from YouTube, TikTok, Bilibili, Apple Podcasts, SoundCloud, or any of 30+ other platforms, or drag in an mp3, mp4, wav, or plain text file. When a video already carries subtitles, the transcript arrives in seconds with no audio download at all, and Whisper takes over only when captions are missing. A badge shows which path each job took, and live progress streams to the page as the job runs. Summaries come out in 11 languages, and a Translation tab appears whenever the transcript language differs from the one you picked. The original video or audio stays beside the results with an inline player and a one-click download, and every tab has its own download icon. Models are your choice: enter any OpenAI-compatible base URL and key in the interface, press Fetch to list what is available, and pick one. Agents get a command line tool, an MCP server, a Claude Code skill, and a Codex plugin. Running on a dedicated VPS on RepoCloud with guaranteed CPU, RAM, and SSD, full root SSH access, and a browser serial console. Apache-2.0 licensed.
Benefits
- Skip Reading Long Recordings
- Get a structured summary with key takeaways for every video or podcast, so one minute tells you if the full transcript deserves your attention.
- Fast When Subtitles Exist
- Native captions are extracted in seconds without downloading audio, and Whisper speech recognition runs only for sources that lack subtitles, saving time and bandwidth.
- Use Any Language Model
- Point the app at OpenAI, OpenRouter, or a local server through an OpenAI-compatible URL, then choose the model that fits your budget and quality needs.
- Keep Every Source Together
- Transcript, summary, translation, and the original media sit in one results card, each with its own download button for notes, archives, and sharing.
Features
- Link And File Input
- Paste a URL from 30+ platforms or upload mp3, mp4, m4a, wav, webm, mkv, ogg, flac, or txt files for the same processing pipeline.
- Whisper Transcription
- Faster-Whisper converts speech to text when subtitles are missing, with FFmpeg normalizing uploaded media before recognition begins.
- Multilingual Summaries
- Choose from 11 summary languages, and let the app translate the transcript automatically when its language differs from your selected output.
- Live Progress Badges
- Server-sent events report each stage as it happens, with a badge that marks each job as using subtitles or Whisper.
- Original Media Download
- Keep the source video at up to 720p alongside the text, preview it inline, and save it with one click or switch it off.
- Agent Integrations
- Run the same pipeline from a command line tool, a stdio MCP server, a Claude Code skill, or a Codex plugin.