Notes And Limitations

What this page covers

Deployment-aware constraints you should expect when planning usage and scale.

Core workflow

1. Plan for personal or small-team load by default.
2. Confirm embedding and extraction model strategy for cost/performance.
3. Decide when optional LLM extraction should be enabled.
4. Verify gpg availability if you need encrypted backups.

Notes

• Postgram is optimized for personal and small-team scale.
• Compose chooses OpenAI embeddings when given a real key; otherwise it uses Ollama for the local path.
• LLM extraction is optional and off by default.
• Backup encryption requires gpg.