Ollama (local) logo
Ollama (local)
Ollama (local) is available as a provider in Qorven. Runs locally — no outbound calls, no API costs.

​
At a glance

Categoryopenai-compat
API basehttp://localhost:11434/v1
Key env varnone (local)
Provider typeOpenAI-compatible

​
Setup

1

Get credentials from Ollama (local)

Install Ollama (local) locally. See their docs.
2

Add the key in Qorven

Settings → Provider Keys → Add provider → Ollama (local) → paste key → Save.The key is AES-256-GCM encrypted with your install’s encryption key before it hits the database.
3

Mark a default model

Models Hub → Models → pick which Ollama (local) model is the default for new Qors.
4

Test

Ask Prime any question. In the LLM trace you’ll see provider=ollama.
The full model list is in the models catalog. Qorven’s catalog keeps costs + context windows + capability flags up to date weekly from the providers.

​
Rate limits & failover

Qorven supports multiple keys per provider. When one rate-limits, the next is tried automatically. When all keys on a provider are exhausted, failover moves to the next provider in your priority list. Failover →

​
Common issues

Key is wrong, expired, or from the wrong org. Test with qorven providers verify ollama.
Provider is throttling. Add another key, or lower your Qor’s model to a less-busy tier. Failover handles this automatically if configured.
Usually a missed top-up. Check the provider’s billing dashboard. Qorven will fail over to the next provider; flag loudly if all are exhausted.
Check your QORVEN_OUTBOUND_TIMEOUT (default 120s). Some local models need more. For Ollama (local) specifically: check the local daemon is up and the port is open.
Logo displayed under nominative fair use. See trademarks & attribution for owner, source library, and removal requests.

Provider keys

Full key management UI + CLI.

Models overview

Catalog, defaults, cost.

Failover

Multi-key, multi-provider rotation.

qorven providers CLI

Scripted provider management.