Content Distribution Agent (MCP-style) — Python Backend (FastAPI)

This is a working backend that follows the MCP-style tool pattern described in the design doc. It exposes a small set of agent tools via HTTP, runs an async worker/scheduler, and includes a mock Instagram connector. It uses SQLite for storage and is self-contained — no external services required.

Features in this MVP

• MCP-like tools (HTTP endpoints) the agent (or you) can call:
  • repurpose_reel_to_post — turn a "reel transcript" into a cross-platform post + hashtags.
  • schedule_post — persist a scheduled post and enqueue a publishing task at the given time.
  • publish_reel — immediately publish via a platform connector (mock Instagram here).
  • fetch_engagement_metrics — returns fake metrics and stores analytics.
• Agent — a tiny heuristic caption/hashtag generator with A/B option.
• Scheduler & Worker — in-process asyncio scheduler + queue; safe for single-instance demos.
• Connectors — base class + Instagram mock (logs instead of calling real APIs).
• Storage — SQLite with SQLAlchemy models for posts, schedules, analytics.
• FastAPI — OpenAPI docs at /docs for quick testing.

Notes

• This MVP avoids real platform APIs so it runs anywhere. Swap the mock connector with real implementations later.
• For multi-instance production, replace the scheduler with a durable queue (Redis/Rabbit/Kafka) and a separate worker process.

Quickstart

1) Local (Python 3.10+)

python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install -r requirements.txt
uvicorn app.main:app --reload

Open http://127.0.0.1:8000/docs to try endpoints.

2) Docker

docker build -t content-agent .
docker run -p 8000:8000 content-agent

Key Endpoints

• POST /tools/repurpose_reel_to_post
• POST /tools/schedule_post
• POST /tools/publish_reel
• GET /tools/fetch_engagement_metrics
• GET /health

Minimal Flow

1. Call POST /tools/repurpose_reel_to_post with a transcript to get platform-optimized text.
2. Call POST /tools/schedule_post (or publish_reel) with the content.
3. The in-process scheduler triggers the worker which calls the mock connector.
4. Check /tools/fetch_engagement_metrics?post_id=... for (fake) analytics.

Swapping Connectors

• Implement PlatformConnector for platforms (Instagram, TikTok, YouTube).
• Register in app/services/registry.py.

Project Layout

app/
  main.py                 # FastAPI app, background workers, routes
  db.py                   # SQLite engine + session
  models/                 # SQLAlchemy models
  connectors/             # Base + Instagram mock
  services/
    registry.py           # MCP-like tool registry
    scheduler.py          # Async scheduler + worker
    agent.py              # Simple caption/hashtag logic
    media.py              # (stub) media helpers
    analytics.py          # Analytics collection helpers
  utils/time.py           # helpers for timezone-aware scheduling
requirements.txt
Dockerfile
README.md

License

MIT — do whatever you want, no warranty.