AI-based Reddit post aggregator 🚀

Scout reads subreddits for you and extracts data from "interesting" posts based on your preferences.

First of all, you need to obtain Reddit and Gemini API credentials.

Reddit API credentials are optional, you can use anonymous mode but it will have lower rate limits.

Create a .env.docker file in the root of the project and fill it with your data.

Template:

POSTGRES_CONN_STRING=postgres://scout:super_secret_password@postgres:5432/scout

GEMINI_API_KEY="<your-gemini-api-key>"

REDDIT_CLIENT_ID="<your-reddit-client-id> or omit to use anonymous mode"
REDDIT_CLIENT_SECRET="<your-reddit-client-secret> or omit to use anonymous mode"
REDDIT_USERNAME="<your-reddit-username> or omit to use anonymous mode"
REDDIT_PASSWORD="<your-reddit-password> or omit to use anonymous mode"
REDDIT_USER_AGENT="<your-reddit-user-agent, e.g. scout/1.0> or omit to use anonymous mode"

If want to use another model or scraping settings, you can do it in settings.yaml

Run docker compose up in the root of the project.

Open UI at localhost:5602.

You can add tests to your profiles to ensure that relevancy prompts work correctly!

Scout consists of the following components:

• postgres - PostgreSQL database for storing posts, analysis results and metadata
• reddit provider - Service for interacting with Reddit API (scrapes posts from subreddits, enriches them and schedules them for analysis)
• analysis tasks queue - Postgres-based queue for asynchronous processing of posts
• analyzer - Service that runs analysis of posts and saves results to the database
• api - HTTP API for interacting with analyzer and reddit-provider
• ui - Frontend application

MIT