Adding Providers
Providers are managed in Configuration > Model Providers. The installer can pre-seed one cloud provider, but the dashboard is the normal place to add, edit, disable, and test providers after installation.
Installer Quick Choices
During install, ClawNex can configure one of these providers:
| Provider | Notes |
|---|---|
| OpenRouter | Multi-model cloud provider |
| Anthropic | Claude models through the Anthropic API |
| OpenAI | GPT models through the OpenAI API |
| NVIDIA NIM | NVIDIA-hosted API endpoint; keys are available from NVIDIA’s build portal |
| Skip | Configure providers later in the dashboard |
NVIDIA NIM uses https://integrate.api.nvidia.com/v1 by default. Start at build.nvidia.com/models to create an API key and select a model.
Dashboard Provider Catalog
The dashboard currently supports:
- Local: LM Studio, Ollama
- Cloud APIs: OpenRouter, Anthropic, OpenAI, Google Gemini, Azure OpenAI, Groq, Together AI, Mistral AI, Cohere, NVIDIA NIM, Fireworks AI, DeepSeek, Perplexity
- Generic: OpenAI Compatible
Add a Provider
Open Model Providers
Go to Configuration > Model Providers and click Add Provider.
Pick the provider type
Choose the closest provider type. Use OpenAI Compatible for endpoints that expose the OpenAI-compatible chat/completions API.
Fill connection settings
Cloud providers need an API key. Local providers usually do not.
Common local defaults:
| Provider | API base URL |
|---|---|
| LM Studio | http://localhost:1234/v1 |
| Ollama | http://localhost:11434/v1 |
Common cloud defaults:
| Provider | API base URL |
|---|---|
| OpenRouter | https://openrouter.ai/api/v1 |
| Anthropic | https://api.anthropic.com/v1 |
| OpenAI | https://api.openai.com/v1 |
| NVIDIA NIM | https://integrate.api.nvidia.com/v1 |
Save and restart LiteLLM when prompted
Saving a provider syncs litellm/config.yaml. If the dashboard shows a restart-required banner, restart LiteLLM from the dashboard or with the service manager.
Verify the Provider
- Open Infrastructure and confirm the LiteLLM proxy is online.
- Open Shield Tests and run the validation suite.
- Send a test message through an OpenClaw agent and confirm traffic appears in Traffic Monitor.
Provider API keys are stored server-side and masked in GET responses. Never paste production keys into screenshots, logs, or support tickets.