Configure the provider

Run these in the AI Labs terminal

Run hrida-agent-sdk in AI Labs' built-in terminal so the provider is saved on the machine AI Labs runs on. Install it first: see hrida-agent-sdk.

The provider is the model service hrida-agent-sdk talks to. Set it once with the /provider command; the choice is saved and reused every time hrida-agent-sdk starts, including when AI Labs runs it as an agent.

1. Open the provider setup

Start hrida-agent-sdk, then type:

/provider

The Set up a provider profile screen shows your current provider, model, endpoint, and saved profile.

2. Choose a provider

Use the arrow keys and press Enter:

Option	Use it for	You'll be asked for
Auto	Use local Ollama if it's running, otherwise set up an OpenAI-compatible provider	A goal (Balanced, Coding, or Latency)
Anthropic	Claude models	Anthropic API key (sk-ant-...), model name
Ollama	A local or self-hosted Ollama model, no API key	Ollama URL, optional API key, model
OpenAI-compatible	Hrida.ai, OpenAI, DeepSeek, OpenRouter, Groq, LM Studio, and similar APIs	API key, base URL, model name
Gemini	Google Gemini models	Gemini API key (AIza...), model name
Codex	Existing ChatGPT Codex CLI login	A Codex profile
NVIDIA NIM	NVIDIA-hosted models (Kimi K2, Llama, DeepSeek, Qwen, Mistral, and more)	NVIDIA API key (nvapi-...), base URL, model
Clear saved profile	Remove the saved profile and go back to normal startup	Nothing (only shown when a profile is saved)

For a field that shows a default, leave it blank and press Enter to accept the default.

3. Example: connect to Hrida.ai

Choose OpenAI-compatible and enter:

Step	Prompt	Enter
1 of 3	API key	Your Hrida.ai API key
2 of 3	Base URL	https://api.hrida.ai/hridaai/v1
3 of 3	Model name	HridaAI

4. Restart and check

Exit hrida-agent-sdk (/exit) and start it again. The startup banner should show your choice:

Provider  OpenAI
Model     HridaAI
Endpoint  https://api.hrida.ai/hridaai/v1
● cloud   Ready — type /help to begin

The profile is saved to ~/.hrida-agent-sdk/provider-profile.json. AI Labs reads the same file, so chats that use the hrida-agent-sdk agent pick up the new provider and model automatically.

API keys

The profile file stores your API key. Don't share or commit it. To switch providers later, run /provider again; to remove the saved key, choose Clear saved profile.

Troubleshooting

• The banner still shows the old provider — restart hrida-agent-sdk; the profile is only read at startup. Provider environment variables set in the shell (such as OPENAI_API_KEY or HRIDA_AGENT_SDK_CODE_USE_OPENAI) take priority over the saved profile.
• 401 Unauthorized when chatting — the API key is wrong or expired. Run /provider and enter it again.
• 404 or model not found — check the base URL and model name are spelled exactly as your provider documents them.
• Ollama: Could not connect — make sure Ollama is running and reachable at the URL you entered. When AI Labs runs in Docker, localhost means the container, not your PC; use http://host.docker.internal:11434 instead.