Configure custom AI model providers — Anthropic, OpenAI, Google, or self-hosted endpoints.

Bring Your Own Model (BYOM)

QANATIX uses AI models for features like chat, description generation, and data extraction. By default, QANATIX uses its own managed model. With BYOM, you can configure your own model provider.

Supported providers

Provider	Models	Config key
Anthropic	Claude 4 Opus, Claude 4 Sonnet, Claude 3.5 Haiku	`anthropic`
OpenAI	GPT-4o, GPT-4o-mini, o3-mini	`openai`
Google	Gemini 2.5 Pro, Gemini 2.5 Flash	`google`
Custom endpoint	Any OpenAI-compatible API	`custom`

Configure via dashboard

Go to Settings > Models in the QANATIX dashboard
Select a provider
Enter your API key
Choose a default model
Save

The API key is encrypted at rest and never exposed after saving.

Configure via API

curl -X PATCH https://api.qanatix.com/api/v1/portal/account \
  -H "Authorization: Bearer <JWT>" \
  -H "Content-Type: application/json" \
  -d '{
    "model_provider": "anthropic",
    "model_api_key": "sk-ant-...",
    "model_name": "claude-sonnet-4-20250514"
  }'

Parameters

Field	Type	Required	Description
`model_provider`	string	yes	`anthropic`, `openai`, `google`, or `custom`
`model_api_key`	string	yes	API key for the provider
`model_name`	string	no	Specific model ID. Falls back to provider default if omitted.
`model_endpoint`	string	no	Custom endpoint URL (required for `custom` provider)

Custom model endpoints

For self-hosted or third-party OpenAI-compatible APIs (e.g., vLLM, Ollama, Azure OpenAI):

curl -X PATCH https://api.qanatix.com/api/v1/portal/account \
  -H "Authorization: Bearer <JWT>" \
  -H "Content-Type: application/json" \
  -d '{
    "model_provider": "custom",
    "model_api_key": "your-key",
    "model_name": "my-model",
    "model_endpoint": "https://llm.internal.company.com/v1"
  }'

The endpoint must be OpenAI-compatible (/v1/chat/completions).

Default model selection

If no BYOM is configured:

Cloud (managed): QANATIX uses its own managed model — no API key needed from you.
Self-hosted: You must configure a model provider. QANATIX cannot make external API calls from air-gapped deployments without an explicit provider configuration.

Self-hosted vs cloud

Behavior	Cloud	Self-hosted
Default model	Managed by QANATIX	Must configure BYOM
API key storage	Encrypted in managed DB	Encrypted in your Postgres
External calls	QANATIX proxy	Direct from your infra
Air-gapped support	N/A	Use `custom` provider pointing to local endpoint

Verify configuration

After configuring, test with a chat query:

curl -X POST https://api.qanatix.com/api/v1/chat \
  -H "Authorization: Bearer sk_live_..." \
  -H "Content-Type: application/json" \
  -d '{"message": "How many records do I have?"}'

If the model provider is misconfigured, you will get a 502 error with details about the upstream failure.

Bring Your Own Model

On this page