Configuration Guide
OpenHuman Auto-Fetch Guide — Automate Your Entire Digital Life
2026-05-25~7 min read
Auto-Fetch is OpenHuman's automatic data synchronization system. Once configured, it continuously pulls data from your connected services — email, calendar, code repositories, and messaging — into your Memory Tree. Your AI always has the latest context.
What Auto-Fetch Does
- Periodic sync — automatically fetches new data from all integrations
- Intelligent scheduling — adapts sync frequency based on how active each service is
- Incremental updates — only fetches what changed since last sync
- Background operation — runs silently without interrupting your work
Setting Up Auto-Fetch
Auto-Fetch is enabled by default once you connect any integration. To configure it:
[auto_fetch]
enabled = true
global_interval_minutes = 15
# Per-service overrides
[auto_fetch.services]
gmail = { interval_minutes = 10, enabled = true }
github = { interval_minutes = 30, enabled = true }
slack = { interval_minutes = 5, enabled = true }
calendar = { interval_minutes = 60, enabled = true }
notion = { interval_minutes = 60, enabled = true }Smart Sync
Auto-Fetch uses smart sync to minimize resource usage:
- Only active services are polled frequently
- Services without new activity are polled less often (backoff)
- During your focused hours, sync frequency decreases automatically
- Rate limits are respected to avoid API throttling
What You Can Do with Auto-Fetch
With Auto-Fetch running, your AI stays constantly up to date:
- "What new emails arrived while I was in my meeting?"
- "Did any of my PRs get reviewed since this morning?"
- "Summarize everything that happened in the past hour"
- "What changed in my calendar for next week?"
Performance Considerations
- CPU usage — minimal, sync runs in a separate process
- RAM — synced data is stored in Memory Tree; set limits in config
- API costs — Auto-Fetch itself doesn't use API calls; only the LLM queries do
Disabling Auto-Fetch
If you prefer manual control, disable Auto-Fetch globally or per-service:
[auto_fetch] enabled = false # Disable all auto-syncIntegrations still work — you can manually trigger a fetch from the settings panel.