Bring your own API key
Step 4 asks: "Which models should power this namespace?" Choose Bring your own API key to use your own account with a commercial model provider instead of, or alongside, Scirix-managed models.
Agent pools🔗
If your organization already owns one or more agent pools, the form first shows an Agent pool picker with two options:
- Attach to your existing pool. Reuse a pool you already have. If you have more than one, a dropdown lets you choose.
- Create a new pool instead. Provision a fresh private pool.
If this is your first time, no picker appears and a new pool is created for you automatically.
Supported backends🔗
Backends with ready-made model presets:
| Backend | Example presets |
|---|---|
| Anthropic | Claude Opus 4.8 (recommended default), Fable 5, Sonnet 5, Haiku 4.5 |
| OpenAI | GPT-5.6 Sol / Terra / Luna, GPT-5.4 Mini / Nano, text-embedding-3-large / small |
| Mistral | Mistral models |
| Kimi (Moonshot) | Moonshot models |
| Gemini | Google Gemini models |
| xAI | Grok models |
| Cohere | Cohere models |
| Databricks | Requires your own server URL as API base |
Backends available in Custom mode only, with no presets:
| Backend | Notes |
|---|---|
| Azure OpenAI | Requires your resource URL as API base, for example https://<resource>.openai.azure.com |
| AWS Bedrock | Custom configuration |
| Ollama (self-hosted) | API base required, for example http://<host>:11434. API key optional. |
| vLLM (self-hosted) | API base required, for example http://<host>:8000/v1. API key optional. |
The form opens with OpenAI and its top preset selected by default.
Adding your provider🔗
- In Step 4, select "Bring your own API key".
- Fill in the form.
| Field | What to enter |
|---|---|
| Name | A display name, for example acme/claude-sonnet. |
| Backend | The provider, from the list above. |
| Model preset | A ready-made model configuration, or Custom to define your own. |
| Model | Only shown for Custom. The model identifier, for example anthropic/claude-sonnet-5. |
| API base | The server URL. Optional for most backends; required for Azure, Databricks, Ollama and vLLM. A hint shows the expected format. |
| API key | Your provider API key, for example sk-…. Stored encrypted and never shown again. |
| Advanced settings | Custom preset only: capabilities, rate limits, context window and pricing. See Capabilities. |
| Provider groups | Pick at least one group for this provider to join. |
| Agent pool | Attach to an existing pool or create a new one. Shown only if you already own pools. |
- Click "Add provider".
Validation to be aware of:
- "An API key is required." Every backend needs a key except Ollama and vLLM.
- "Assign the provider to at least one group." At least one provider group is mandatory.
- Backends that require an API base show: "[Backend] providers need the server URL as API base."
- In Advanced settings: rate limit, concurrency and token limits must be positive whole numbers; costs must not be negative.
What happens next: adding the provider provisions your organization's private agent pool, a dedicated indexer and auditor, and moves the namespace onto it. Presets lock in the right capabilities, token limits and pricing automatically. Custom uses whatever you set in Advanced settings.
How your provider gets used🔗
Your provider joins one or more provider groups. Scirix roles, the tasks described on the Capabilities page, resolve their model through these groups. Your own providers in a group are preferred over the platform's. That is how your key takes over work for your namespace.
After onboarding, you can manage providers and groups under Settings, then LLM providers.
Step 5 for your own API key🔗
Continue with The embedding index to choose which model builds the search index.