CA

chengyi-ai/douyin-image-post-scheduler

Data science & ML

This page was translated automatically. Technical details may differ from the original. View the original

108 stars Quality 61 Trend 29MIT

Codex Skill for batching, scheduling, and verifying Douyin image posts.

Overview

An open-source skill that can be installed on Codex/compatible agents: generates reproducible TikTok text queues from structured local creatives, and completes batch uploads, valid topic selection, music settings, scheduled publishing, and backend verification in Chrome where users are logged in. This project is not an official tool for Douyin. The webpage structure may change; you must check the current page and account status before publishing. - Batch check whether graphic and text materials are complete; - Generate posting titles up to 20 characters per first image; - Reserve space for 1000-character main text and 5 valid topics; - Multiple posts per day, fixed time windows, fixed random seeds, replayable scheduling; - Avoid misinterpreting plain text #标签 as valid topics; - Verify images, titles, main text, topics, music, and time before each post; - Continue from the last successful entry after interruption to avoid duplicate postings; - Distinguish between "Under Review" and "Scheduled Release". Call $skill-installer in Codex to install the following Skill directory: After reopening the Codex task, you can use the $douyin-image-post-scheduler. The script only reads local material and generates a project file, without opening the browser or publishing content. The seed in the output can be used to keep the same batch of schedules exactly consistent. 1. Check the accurate material catalog and current Douyin account; 2. Generate JSON queues and Markdown schedules; 3. Upload 8 images and fill in titles and body text; 4. Place 5 tags in separate rows and activate them one by one from the candidate panel; 5. Select the first song in your favorites; 6. Check for publicity, save allowed, no cross-platform synchronization, and scheduled time; 7. Submit after receiving batch release confirmation; 8. Return to the content management page for each item, verify the total number and last item after the batch ends. Actual external release requires user confirmation. This skill does not store account passwords, read browser cookies, or bypass verification codes or login verification.

Overview (machine translated)

An open-source skill that can be installed on Codex/compatible agents: generates reproducible TikTok text queues from structured local creatives, and completes batch uploads, valid topic selection, music settings, scheduled publishing, and backend verification in Chrome where users are logged in.

This project is not an official tool for Douyin. The webpage structure may change; you must check the current page and account status before publishing.

What problem to solve

  • Batch check whether images and text materials are complete;
  • Generate a posting title of up to 20 characters based on the first image;
  • Reserve space for a 1000-word main text limit and 5 valid topics;
  • Multiple daily reproduction schedules with specified time windows and fixed random seeds;
  • Avoid misclassifying plain text #标签 as valid topics;
  • Verify images, titles, body text, topics, music, and time before each post;
  • After interruption, continue from the last successful entry to avoid duplicate publications;
  • Distinguish between “Under Review” and “Scheduled Release.”

Warehouse Structure

The warehouse does not contain Douyin accounts, cookies, original materials, published content, scheduling results, or screen recordings.

Maintenance files do not belong to Skill: CLAUDE.md (project agreement for the Agent), tools/ (warehouse inspection, recording feedback), tests/, docs/, .github/ (CI and feedback processing workflow).

Installation

Codex Skill Installer

Call $skill-installer in Codex to install the following Skill directory:

Manual installation

After reopening the Codex task, you can use $douyin-image-post-scheduler.

Material format

Each item is placed in a table of contents starting with a three-digit number:

By default, each requires 8 images and at least 5 #话题. These values can be adjusted via script parameters.

The gentleman forms a queue

The script only reads local material and generates a schedule file, without opening the browser or publishing content. The seed in the output can be used to keep the same batch of schedules exactly consistent.

Used in Codex

The default Skill release process is:

  1. Check the accurate material directory and current Douyin account;
  2. Generate JSON queues and Markdown schedules;
  3. Upload 8 images and fill in the title and main text;
  4. Place 5 tabs in separate rows and activate them one by one from the candidate panel;
  5. Select the first song in your favorites;
  6. Check for disclosure, allow retention, avoid cross-platform synchronization, and schedule timelines;
  7. Submit after obtaining batch release confirmation;
  8. Return to the content management page for each entry, and verify the total number and last item after the batch ends.

Actual external release requires user confirmation. This skill does not store account passwords, read browser cookies, or bypass verification codes or login verification.

Verification

Test coverage queue generation, fixed random seeds, image/tag/title constraints, daily entries, and minimum time interval. tools/validate_repo.py Check Skill metadata, paths in documents, redline declarations, and the absence of account credentials, materials, and scheduling results in the repository.

Feedback

This page was translated automatically. Technical details may differ from the original.

Recommended Tools

Try a different keyword or remove a filter.

Install

npx skillfish add chengyi-ai/douyin-image-post-scheduler
New to MCP? Start with the mcpskill.org guide →