Analysis updated 2026-05-18
Automatically save videos you forward to yourself on Douyin into a searchable Feishu database.
Get an automatic transcript and summary of a saved Douyin video's spoken content.
Build a personal knowledge base of interesting videos without manual copy-pasting.
Manually transcribe a single downloaded video or audio file using Whisper.
| dannyzz2/douyin-share | albertaworlds/japanese-text-cleaner | ayangabryl/ngx-digit-flow | |
|---|---|---|---|
| Stars | 30 | 30 | 30 |
| Language | TypeScript | TypeScript | TypeScript |
| Setup difficulty | hard | easy | easy |
| Complexity | 4/5 | 2/5 | 2/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Node.js, Python, FFmpeg, an authenticated Feishu CLI, and a real (hidden) browser window to avoid verification checks.
Douyin Self-Share Capture watches a person's own Douyin inbox on the web version of the app and automatically copies anything they share to themselves into a Feishu Base, which is a spreadsheet-like database tool used widely in China. The idea is that when someone forwards a video to their own account from a phone, from Douyin's desktop app, or from the web, this tool notices the new message and turns it into a saved, searchable record, effectively building a personal knowledge base out of videos a person wanted to keep. The tool runs on a computer using Playwright, a browser automation tool, with a persistent browser profile so a login only needs to happen once. When a shared video appears, it clicks the card to find the video's unique ID, then reads and cleans up the caption, removing things like author badges and publish times so only the real title text remains. Once that basic record is saved, a background process takes over: it fetches the exact video file, checks that the audio track is valid, transcribes the video's speech to text using OpenAI's Whisper speech recognition model, converts any Traditional Chinese text to Simplified Chinese, generates a short summary, and writes everything back into the correct row in Feishu. The Feishu record includes fields for title, content type, original link, original text, author or source, summary, capture time, processing status, a deduplication ID, any error message, and a link back into the user's knowledge base. Two additional fields for key points and tags are mentioned as not yet part of this early version. Setup requires Node.js, Python 3.10 or newer, FFmpeg, and a Feishu command-line tool already authenticated with write access to the target table. Summaries can be generated with a mock mode that skips any external AI call, or through an OpenAI-compatible provider such as Agnes AI. The project deliberately avoids fully headless browser mode because it can trigger Douyin's verification pages, instead running a real but hidden browser window. Sensitive files like the environment config, browser profile, and downloaded media are excluded from version control by default. The README does not state a license.
A tool that watches your own Douyin self-shared videos, transcribes them with Whisper, and saves organized records into a Feishu Base as a personal knowledge library.
Mainly TypeScript. The stack also includes TypeScript, Playwright, Python.
Setup difficulty is rated hard, with roughly 1h+ to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.