This is a fully autonomous, AI-driven content and outreach system that:
- Writes long-form SEO blog posts from keyword queues
- Adds images, YouTube videos, citations, and internal links
- Publishes the post via API
- Then researches 100+ relevant websites, finds emails, writes personalized pitches, and sends outreach campaigns — all automatically.
Built using a custom stack of LangChain agents, AI tool integrations, FastAPI, and SendGrid.
- Pops keywords from a queue and runs deep research via Serper, YouTube API, and more
- Generates long-form blog content using OpenAI + Google Gemini
- Adds:
- Internal links (based on site structure)
- Relevant images (via Imagen or scraped assets)
- Embedded YouTube videos
- Citations for factual claims
- SEO optimization included: metadata, keyword targeting, readability
- AI agent reads each blog post and identifies 100+ relevant websites to pitch
- Evaluates site quality and audience match using AI
- Finds real email addresses using web scraping and pattern recognition
- Writes uniquely personalized emails per site (no templates)
- Compliments their writing
- Justifies the blog post’s value to their audience
- Sends via SendGrid and tracks delivery + responses
- OpenAI GPT (LangChain agents)
- Google Gemini (factual validation + content synthesis)
- Google Imagen (optional image generation)
- LangChain with multi-agent orchestration
- Serper API (for web search)
- YouTube Data API (video embedding)
- SendGrid (outbound email)
- Google Cloud (LLMs, storage)
- FastAPI (Python)
- Uvicorn (ASGI server)
- REST API endpoints for orchestration
- Modular
src/architecture
- MySQL (MariaDB)
- Used to store:
- Prospective outreach URLs
- Keyword queue
- Blog metadata
- Email delivery stats
- Dockerized
- Vercel (optional serverless deployment)
- Render support (via
render.yaml) .envbased secrets handling
| File | Purpose |
|---|---|
main.py |
FastAPI app entry point |
src/ContentAPIHandler.py |
Blog writer + optimizer agent |
src/ProspectGenerator.py |
Backlink research & outreach |
src/EmailReplyProcessor.py |
Email tracking + response handling |
src/DatabaseService.py |
DB connection + query methods |
local_tests.py |
Manual CLI test runner |
render.yaml / vercel.json |
Hosting configs |
- Fully autonomous from keyword → published post → AI outreach
- No templates: each pitch is written by an LLM for that specific site/contact
- AI agents choose the strategy dynamically based on blog content
- Real-world infrastructure: emails are delivered, links go live, stats are tracked
- API keys are secured via
.env - SendGrid and DB credentials handled securely
- Never commit sensitive keys — system uses dynamic environment config
- ✅ Content generation working
- ✅ Outreach agent functioning
- ✅ Posts published and live
- ✅ Email tracking implemented