Models
Configure and use your own models
Orqra uses the Provider connections you configure on this device. Complete these steps once, then choose an enabled, compatible model wherever a workflow needs it.
What you will complete
Set project defaults or choose a compatible enabled model on a processing node. Review Request Preview, then run only when the inputs, model, and parameters are correct.
In this guide
- 1Open Models — Open Models and choose the Provider connection you want to configure. You can use a supported Provider or a custom compatible endpoint.
- 2Save credentials locally — Enter the API Key or Provider-specific credentials, then save. Save sensitive changes before testing or syncing models; a saved API Key is not displayed again.
- 3Add and enable models — Add the models you need to My Models and enable them. Free supports up to 3 enabled models; Personal has no edition-level enabled-model limit.
Model and connection support
Separate the connection method, the API protocol, and the exact models verified by Orqra before choosing a model for a workflow.
OpenRouter connection
Connect with your own OpenRouter API key, sync its model catalog, and add models to My Models. Orqra does not provide or resell model credits.
OpenAI-compatible API
Connect a third-party service, gateway, or self-hosted endpoint with a custom Base URL. Orqra supports Chat Completions at /chat/completions and Responses at /responses; actual compatibility depends on the endpoint implementation.
Models verified by Orqra
These exact OpenRouter model IDs have completed Orqra capability adaptation and real-call verification for the documented v1 support scope.
| Capability | Model | Exact model ID |
|---|---|---|
| Text | GPT-5.6 Sol | openai/gpt-5.6-sol |
| Text | GPT-5.6 Luna | openai/gpt-5.6-luna |
| Text | Claude Opus 5 | anthropic/claude-opus-5 |
| Text | Claude Sonnet 5 | anthropic/claude-sonnet-5 |
| Text | Gemini 3.6 Flash | google/gemini-3.6-flash |
| Text | Gemini 3.5 Flash Lite | google/gemini-3.5-flash-lite |
| Text | Kimi K3 | moonshotai/kimi-k3 |
| Text | Kimi K2.5 | moonshotai/kimi-k2.5 |
| Text | DeepSeek V4 Pro | deepseek/deepseek-v4-pro |
| Text | DeepSeek V4 Flash | deepseek/deepseek-v4-flash |
| Image | Nano Banana Pro | google/gemini-3-pro-image |
| Image | Nano Banana 2 | google/gemini-3.1-flash-image |
| Image | GPT Image 2 | openai/gpt-image-2 |
| Video | Veo 3.1 | google/veo-3.1 |
| Video | Veo 3.1 Fast | google/veo-3.1-fast |
| Video | Seedance 2.0 | bytedance/seedance-2.0 |
| Video | Seedance 2.0 Fast | bytedance/seedance-2.0-fast |


Before you start
- Your API Key is stored locally. Never send it to Orqra support or enter it on the Orqra website.
- Use the endpoint and credential format documented by your Provider or custom gateway.
- Model availability, capabilities, rate limits, and fees are controlled by the Provider. Orqra does not include AI credits.
Complete the task
- 1
Models
Open Models
Open Models and choose the Provider connection you want to configure. You can use a supported Provider or a custom compatible endpoint.
Expected result
The Models page shows the connection you want to configure.
- 2
Service Connections
Configure the connection
Confirm the Provider type and required connection fields. Change the Base URL only when the Provider or your own gateway requires it.
Expected result
The Provider type, Base URL when required, and Provider-specific fields match your Provider account.
- 3
Save
Save credentials locally
Enter the API Key or Provider-specific credentials, then save. Save sensitive changes before testing or syncing models; a saved API Key is not displayed again.
Expected result
The connection is saved locally and the saved API Key value is no longer displayed.
- 4
Test Connection / Sync Models
Test and sync
Use Test Connection, then Sync Models to refresh the available list. These actions contact the Provider, but they do not run a generation request.
Expected result
The connection result is visible and the Provider model list has been refreshed.
- 5
My Models
Add and enable models
Add the models you need to My Models and enable them. Free supports up to 3 enabled models; Personal has no edition-level enabled-model limit.
Expected result
The models you need appear in My Models and the required models are enabled.
- 6
Project / Processing node
Choose, review, and run
Set project defaults or choose a compatible enabled model on a processing node. Review Request Preview, then run only when the inputs, model, and parameters are correct.
Expected result
A compatible enabled model is selectable as a project default or on the processing node.
Checks before you retry
- Use the endpoint and credential format documented by your Provider or custom gateway.
- Support can explain Orqra BYOK setup fields, but Orqra does not include AI model API credits and does not resell third-party AI APIs. Provider API keys, billing, rate limits, and model availability remain with your own provider.
- Add the models you need to My Models and enable them. Free supports up to 3 enabled models; Personal has no edition-level enabled-model limit.
