Ollama
Run chat, vision and model discovery against a local or hosted Ollama-compatible API.
Capabilities
| Capability | Identifier |
|---|---|
| Chat completion | ai.chat |
| Vision messages | ai.vision |
| Model discovery | ai.models |
Configuration
| Field | Type | Required | Notes |
|---|---|---|---|
| Server URL | url | Yes | Base URL of the Ollama API, without /api. |
| API key secret | password | No | Sent as a Bearer token for hosted Ollama-compatible services. |
| Default model | text | Yes | Default: llama3.2 |
| Timeout | number | Yes | Default: 60 |
About
Connect GrowRig features to any Ollama-compatible API. The built-in runtime supports chat, vision messages containing Ollama image payloads, and model discovery. Create separate instances for local, cloud, or remote GPU servers.
The optional API key is encrypted at rest and is never returned to the browser.
The connection test calls GET /api/version.
Integrations are configured as instances under Control panel → Integrations. Secret fields are AES-GCM encrypted with a local key beside the Grow Core database and are never returned by the API.