Login with TikTok

Auto Upload CLI for TikTok & YouTube Shorts with AI Story Generation & Voiceover

A command-line Python tool that:

1. Generates multi-part suspenseful stories via OpenAI
2. Converts those scripts to voiceovers using ElevenLabs TTS
3. Creates subtitle files and overlays them on background videos via FFmpeg
4. Automatically uploads the finished videos to TikTok and/or YouTube Shorts

This end-to-end pipeline enables you to go from idea to published content with a single script.

---

🔍 Description

auto_upload.py orchestrates the full content creation process:

1. Story generation: calls OpenAI to produce 2–4 part narratives, each ~20–30s long, with cliffhangers.
2. Voiceover: uses ElevenLabs to generate high-quality MP3 narrations for each part.
3. Subtitles: transcribes the TTS output (via Whisper), re-chunks into timed segments, and writes SRT files.
4. Video composition: overlays subtitles on a random background clip using FFmpeg, syncing with the audio.
5. Publishing: uploads the final MP4s to TikTok and/or YouTube Shorts with metadata from a JSON file.

---

⚙️ Features

• AI Story Generation: automatically crafts tense, cliffhanger-driven story scripts.
• Voiceover & Subtitles: integrates ElevenLabs TTS and Whisper for narration plus accurate subtitles.
• Multi-platform Upload: CLI supports TikTok, YouTube Shorts, or both in one run.
• Part-aware metadata: filenames extract part numbers to match JSON entries for captions/tags.
• End-to-end automation: from script to published video, schedulable via cron, GitHub Actions, etc.

---

🛠 Prerequisites

• Python 3.7+
• FFmpeg installed and on your PATH

• Dependencies (install via pip):

  pip install openai python-dotenv elevenlabs requests google-auth-oauthlib google-api-python-client whisper
  
  cd netlify/functions
  npm init -y
  npm install node-fetch@2 dotenv
  

• Environment variables in a .env file or exported:

  • OPENAI_API_KEY – your OpenAI API key
  • ELEVENLABS_API_KEY – your ElevenLabs API key
  • TIKTOK_ACCESS_TOKEN – TikTok OAuth2 token (valid 24h)
  • YOUTUBE_CLIENT_SECRETS_FILE – path to client_secrets.json for YouTube OAuth

---

⚡️ Installation

1. Clone the repo

   git clone https://github.com/yourusername/auto-upload.git
   cd auto-upload
   

2. Install FFmpeg (if not already):

   • macOS: brew install ffmpeg
   • Ubuntu: sudo apt install ffmpeg

3. Install Python deps

   pip install openai python-dotenv elevenlabs requests google-auth-oauthlib google-api-python-client whisper
   

4. Create a .env file in the project root:

   OPENAI_API_KEY=sk-...
   ELEVENLABS_API_KEY=eleven-...
   TIKTOK_ACCESS_TOKEN=...
   YOUTUBE_CLIENT_SECRETS_FILE=client_secrets.json
   

---

🚀 Usage

python auto_upload.py \
  --video /path/to/backgrounds/videoplayback.mp4 \
  --meta /path/to/stories.json \
  --platform both

• --video: path to a background clip directory or single file (will randomize).
• --meta: JSON file for upload metadata (captions, hashtags, tags).
• --platform: tiktok, youtube, or both.

The script will generate story scripts, voiceovers, subtitles, videos, and then upload them, printing API responses.

---

🔐 OAuth Callback Endpoint

TikTok’s OAuth flow requires a publicly accessible HTTPS Redirect URI to receive the authorization code. Simply hosting your code in a public Git repository does not provide this endpoint—you’ll still need one of the following:

Local Server + ngrok

• Run a small Flask (or similar) app on your machine (e.g., on port 8090) to handle /oauth/callback.

• Expose it via ngrok:

  ngrok http 8090
  

• Add the generated https://<your_id>.ngrok.io/oauth/callback URL as your Redirect URI in the TikTok Developer Portal.

---

📅 Scheduling

Automate the entire pipeline via:

• cron (Linux/macOS): e.g. 0 9 * * * python /path/auto_upload.py --…
• Windows Task Scheduler
• GitHub Actions: use a schedule workflow for serverless runs on GitHub’s servers.

---

📝 License

This project is licensed under the MIT License.

---

Ready to generate, voice, and publish? Let’s go!, voice, and publish? Let’s go!*