Troubleshooting

AI health entities are missing

AI health entities are only created when AI is fully configured:

1. Map a camera entity for the grow space.
2. Select a provider (Gemini, OpenAI, or Ollama) — not None.
3. Choose a vision-capable model.

Re-open Options and confirm all three, then reload the integration.

Model discovery failed

If the provider's model list cannot be fetched (network, credentials, or endpoint issues), you can enter a model name manually on the model-selection step. Verify the API key/endpoint and that the account has access to a vision-capable model.

VPD looks wrong

VPD is computed from the air temperature and humidity (canopy), not the water probe. If VPD is off:

• Confirm temperature/humidity are mapped to air/canopy sensors.
• Confirm the water probe is mapped to water_temperature, not the air role.

See Configuration.

Tuya sensors are missing

1. Confirm Tuya is enabled and the access ID, access secret, region, and device IDs are correct.
2. Call tendrilgrow.rebuild_automap to reload and rebuild auto-mapped roles.
3. Check the logs for the custom_components.tendrilgrow logger.

Entities are unavailable after an update

Custom integrations require a Home Assistant restart to load new code. After updating in HACS, restart Home Assistant.

Dashboards reference old entity IDs

If you renamed grow spaces or updated across a version that changed entity IDs, storage-mode (UI-managed) dashboards keep the old IDs. Update the affected cards, or regenerate the dashboard from live entities with scripts/generate_dashboard.py (see Dashboards).

Collecting diagnostics

Download redacted diagnostics from Settings → Devices & Services → TendrilGrow → the entry → ⋯ → Download diagnostics. API keys and the Tuya access secret are redacted, so it is safe to attach to a discussion or issue.

Still stuck?

• Ask in Discussions.
• File a bug report with your Home Assistant version, TendrilGrow version, and sanitized logs.