What it does:
You give it any input — a blog URL, a YouTube video, raw text, or just a topic — and it generates optimized posts for 9 platforms at once: Instagram, Twitter/X, LinkedIn, Facebook, TikTok, Reddit, Pinterest, Twitter threads, and email newsletters.
Each output is tailored to the platform (hashtags for IG, hooks for TikTok, professional tone for LinkedIn, etc.). It also auto-generates images for visual platforms like Instagram, Facebook, and Pinterest,using AI.
Other features:
- Topic Research — scans Google, Reddit, YouTube, and news sources, then uses an LLM to identify trending subtopics before generating content
- Auto-Discover — if you don't even have a topic, it searches what's trending right now (optionally filtered by niche) and picks the hottest one
- Cinematic Ad — upload any photo, pick a style (cinematic, luxury, neon, retro, minimal, natural), and Gemini transforms it into a professional-looking ad
- Multi-LLM support — works with Mistral, Groq, OpenAI, Anthropic, and Gemini
- History — every generation is saved, exportable as CSV
The n8n automation (this is where it gets fun):
I connected the whole thing to an n8n workflow so it runs on autopilot:
1. Schedule Trigger — fires daily (or whatever frequency)
2. Google Sheets — reads a row with a topic (or "auto" to let AI pick a trending topic)
3. HTTP Request — hits my /api/auto-generate endpoint, which auto-detects the input type (URL, YouTube link, topic, or "auto") and generates everything
4. Code node — parses the response and extracts each platform's content
5. Google Drive — uploads generated images
6. Update Sheets — marks the row as done with status and links
The API handles niche filtering too — so if my sheet says the topic is "auto" and the niche column says "AI", it'll specifically find trending AI topics instead of random viral stuff.
Error handling: HTTP Request has retry on fail (2 retries), error outputs route to a separate branch that marks the sheet row as "failed" with the error message, and a global error workflow emails me if anything breaks.
Tech stack:
- FastAPI backend, vanilla JS frontend
- Hosted on Railway
- Google Gemini for image generation and cinematic ads
- HuggingFace FLUX.1 for platform images
- SerpAPI + Reddit + YouTube + NewsAPI for research
- SQLite for history
- n8n for workflow automation
It's not perfect yet — rate limits on free tiers are real — but it's been saving me hours every week. Happy to answer questions.
[link] [comments]
