Daily SEO Feed (Bluesky)

A custom feed generator for Bluesky that surfaces high-quality SEO content based on expert engagement. Built with Python using the AT Protocol SDK.

Features

• Custom algorithm tracking engagement from verified SEO experts
• Weighted scoring system for likes, reposts, and comments
• Time-decay factor for content freshness
• MySQL database for efficient data storage
• Web interface to view current feed content
• Configurable engagement thresholds and weights
• Automatic cleanup of outdated content

Requirements

• Python 3.9+
• MySQL Database
• Bluesky account for feed publication

Setup

1. Clone the repository

2. Install dependencies:

pip install -r requirements.txt

3. Configure environment variables:

   • Copy .env-prod to .env
   • Update the database credentials
   • Set your FEEDGEN_HOSTNAME

4. Initialize the database:

python -m server --rebuild_database

5. Publish your feed:

python publish_feed.py

6. Update your .env with the generated feed URI

Running the Service

There are several ways to run the service based on your needs:

Full Service Mode

Run both the API server and Jetstream processor together:

python -m server

Component Separation

You can run components separately:

API Server Only

Run just the web API and interface:

python -m server --app_only

Jetstream Processor Only

Run just the data processing component:

python -m server --jetstream_only

Development Mode

Run with debug logging enabled:

python -m server --debug

Endpoints

• /.well-known/did.json - DID document
• /xrpc/app.bsky.feed.describeFeedGenerator - Feed metadata
• /xrpc/app.bsky.feed.getFeedSkeleton - Feed content
• /health - Service health check
• / - Web interface showing current feed content

Customization

• Adjust scoring weights in .env
• Modify user_list.txt to update tracked experts
• Configure post lifetime and minimum engagement score

License

MIT