Chapter Promo Builder

A local browser app for turning a novel chapter into a small promo pack:

• three TikTok/Instagram-ready promo images
• three dramatic phrases pulled from the chapter
• captions and short notes for Royal Road and Patreon
• a video helper script that can create a voiceover and a simple vertical MP4 when local tools are available
• a YouTube helper script that randomly selects three background videos from C:\Users\David\Documents\Novels
• daily Instagram/X post packs from C:\Users\David\Documents\Promo Images

Start

Double-click start.bat, then open:

http://127.0.0.1:8765

Paste a chapter, add a title, and choose Build Promo Pack.

For daily social posts, choose the novel abbreviation and day, then choose Create Instagram/X Post. The app uses:

• EN = Eternal Nexus
• HA = Heavenly Ascension System
• SF = Soul Forge Era
• HP = Hundredfold Path

Daily images should follow the existing pattern, such as EN_Monday.png or SF_Friday.png.

The reusable social prompt lives in social_prompt_template.txt, so you can tune the standard wording once and keep using it.

FFmpeg

The app looks for FFmpeg in this order:

1. FFMPEG_EXE from .env.local
2. tools\ffmpeg\bin\ffmpeg.exe
3. any ffmpeg.exe under tools\
4. system PATH

Run this once to install the portable build into tools\ffmpeg:

python setup_ffmpeg.py

If Windows cannot reach the download URL, download ffmpeg-8.1.1-essentials_build.zip from Gyan's FFmpeg builds, put it in tools\, then run setup_ffmpeg.py again.

OpenAI Setup

The app looks for OPENAI_API_KEY in the environment or in .env.local.

When that key is available, it uses ChatGPT for phrases, captions, platform notes, and image prompts. It also tries to generate three images with the OpenAI image API.

When no key is available, it still works: it pulls three strong phrases from the chapter and downloads free placeholder images from an online image source.

Output

Each run creates a folder under campaigns/ with:

• promo-1.png, promo-2.png, promo-3.png
• phrases.txt
• caption.txt
• royal-road-note.txt
• patreon-note.txt
• video-text.txt
• make_video.py
• make_youtube_video.py
• youtube-project-notes.txt

Daily social post runs create a folder under social-posts/ with:

• the selected promo image
• instagram.txt
• x.txt
• alt-text.txt
• prompt-template.txt

Instagram Publishing

The app can publish a generated social post to Instagram through Meta's official Instagram publishing API.

Add these to .env.local:

INSTAGRAM_ACCOUNT_ID=your_instagram_professional_account_id
INSTAGRAM_ACCESS_TOKEN=your_meta_access_token
INSTAGRAM_PUBLIC_BASE_URL=https://your-public-host/social-posts
META_GRAPH_API_VERSION=v25.0

Instagram requires the image to be reachable through a public HTTPS URL. Local files such as C:\...social-posts\...png cannot be posted directly. The app stores the post pack locally, then builds the public URL from INSTAGRAM_PUBLIC_BASE_URL.

After creating a daily post pack, choose Post to Instagram.

For the safer manual flow, choose Manual Instagram Assist. It opens Instagram, opens the generated post folder, and copies the caption to your clipboard. After you publish in Instagram, choose Mark Posted to record completion in the post metadata.

You can also paste the account ID and token into the app's Connections panel. Use Find Account ID after saving a token to look up connected Instagram business accounts.

X Publishing

For the safer manual flow, choose Manual X Assist. It opens X, opens the generated post folder, and copies the X post text to your clipboard. After you publish in X, choose Mark X Posted.

For API posting, save an OAuth 2.0 user access token with posting and media permissions in the Connections panel or in .env.local:

X_ACCESS_TOKEN=your_x_user_access_token

Then choose API Post to X from a generated social post pack.

Buffer Publishing

Buffer can act as the publishing hub for Instagram, TikTok, YouTube, and other connected channels.

In Buffer:

1. Connect your YouTube, Instagram, and TikTok channels.
2. Go to Buffer API settings and create a personal API key.
3. Paste the API key into Connections.
4. Choose Find Buffer Channels to list your connected channel IDs.
5. Paste the channel IDs you want to use into Buffer channel IDs as a comma-separated list.

The current recommended flow is Manual Buffer Assist:

1. Generate a social or TikTok pack.
2. Choose Manual Buffer Assist.
3. The app opens Buffer, opens the local media folder, and copies the caption.
4. In Buffer, use its Google Drive/media picker to attach the media.
5. Paste the copied caption.
6. Choose Mark Buffer Queued in the app.

The Buffer API key is still useful for checking the account/channel connection, but media posting is handled manually because Buffer's API requires public HTTPS media URLs and Cloudflare R2 requires billing setup.

GitHub Code And Media Storage

This folder is intended to sync with:

https://github.com/azureinkblade-ops/Automation-tool

Generated working folders such as campaigns/, social-posts/, tiktok-posts/, and youtube-videos/ stay local by default. Generated media is automatically copied into GitHub Pages storage, committed, and pushed when GITHUB_AUTO_PUBLISH_MEDIA=1:

docs/media/

Enable GitHub Pages for the repository using the main branch and /docs folder. After Pages finishes publishing, media files are available under:

https://azureinkblade-ops.github.io/Automation-tool/media/

If you use a different Pages URL, set it in Connections or .env.local:

GITHUB_REMOTE_URL=https://github.com/azureinkblade-ops/Automation-tool.git
GITHUB_PAGES_MEDIA_BASE_URL=https://your-pages-host/Automation-tool/media
GITHUB_TOKEN=your_fine_grained_token_with_contents_read_write
GITHUB_AUTO_PUBLISH_MEDIA=1

When GITHUB_TOKEN is saved, the app uploads media through the GitHub API and does not depend on local Git authentication or .git folder ownership. The app records public URLs in the generated pack's metadata.json, including public_image_url for Instagram-compatible image posts. Buffer API posting also uses GitHub Pages URLs automatically when Cloudflare R2 is not configured.

TikTok Posting

The app scans:

C:\Users\David\Documents\CloudFlare

It uses image groups like EN_10_1.png, EN_10_2.png, EN_10_3.png, and randomly chooses one sound file from the MP3/WAV files in that same folder.

If the image group for a chapter does not exist yet, enter the chapter number and a TikTok image prompt. The app will generate ABBR_CHAPTER_1, ABBR_CHAPTER_2, and ABBR_CHAPTER_3 into the CloudFlare folder before creating the TikTok pack.

Choose Create TikTok Pack, then:

• Manual TikTok Assist opens TikTok upload, opens the generated folder, and copies the caption.
• Build TikTok Video creates tiktok-video.mp4 from the three images and selected sound using FFmpeg.
• Mark TikTok Posted records completion locally.

Patreon And Royal Road

Chapter promo campaigns include platform-specific notes:

• patreon-note.txt
• royal-road-note.txt

Use the buttons in the chapter promo result:

• Manual Patreon Assist opens Patreon, opens the campaign folder, and copies the Patreon note.
• Mark Patreon Posted records completion locally.
• Manual Royal Road Assist opens Royal Road, opens the campaign folder, and copies the Royal Road note.
• Mark Royal Road Posted records completion locally.

Royal Road does not appear to provide an official public posting API. Patreon has API documentation, but Patreon states developer support for its API is no longer provided, so the app uses the safer manual assistant flow for now.

Posting Schedule

The schedule manager uses:

• Chapter advertising days: Monday, Wednesday, Friday, Sunday
• General promotion days: Tuesday, Thursday, Saturday
• Patreon releases first
• Royal Road releases 14 days after Patreon
• Every scheduled item includes separate Instagram and X copy

The schedule settings live in:

posting_schedule.json

By default, each novel starts at Chapter 1. Edit nextChapter and startDate in that file when you want the planner to match the real chapter queue.

Writing New Chapters

Use Write New Chapter + Build Promo to have ChatGPT draft a chapter from your outline and continuity notes. The app saves:

• the chapter text under chapters/
• summary and continuity notes
• a generated promo campaign under campaigns/

This requires OPENAI_API_KEY in .env.local or saved through the Connections panel.

YouTube Videos From Text

Use YouTube Video to paste chapter text and create:

• youtube-title.txt
• youtube-description.txt
• youtube-tags.txt
• youtube-voiceover.wav
• youtube-video.mp4

The app creates YouTube metadata from the chapter text, picks background videos from C:\Users\David\Documents\Novels, and uses FFmpeg plus Windows speech to build the MP4.

Run make_video.py inside a campaign folder to try creating a voiceover and vertical MP4. It uses Windows speech for audio and ffmpeg for video if available. If you prefer OpenShot, import the generated images and voiceover.wav into a vertical 1080x1920 project.

Run make_youtube_video.py inside a campaign folder to create youtube-video.mp4 from three randomly selected BG_*.mp4 background videos. It uses Windows speech for narration and ffmpeg for the final MP4 when available. If ffmpeg is not installed, use youtube-project-notes.txt to recreate the project in OpenShot with the selected videos, narration, and text overlays.

To point the background picker somewhere else, add this to .env.local:

BACKGROUND_VIDEO_DIR=C:\Users\David\Documents\Novels

Moving Novel Folders

You do not need to move your novels yet. The Scan Examples button looks nearby for prior images, videos, and text files. If you later put examples inside this folder, the scan will find those too.