LLM Configuration#
Agent Master uses a large language model to understand your requests and plan workflows. You choose the provider and supply an API key; the key stays on your machine. Open the configuration from the Settings gear on the command bar.
The Settings Drawer#
Settings: choose a provider, paste an API key, pick a model, and set the active provider and the default figure export.#
LLM Providers — Claude, OpenAI, Gemini, or DeepSeek.
API Key — paste the key for the selected provider; the eye/key toggle reveals it. Save API keys stores it.
Model — the model id to use (for example
claude-sonnet-4-6).Active LLM provider — which configured provider requests actually use.
Default figure export — the format and resolution for figures the agents save (for example
PNG (300 dpi)).
Choosing A Provider And Model#
Pick a provider tab.
Paste its API key.
Set the model id (a sensible default is pre-filled).
Select it under Active LLM provider.
Save API keys.
You can configure more than one provider and switch the active one at any time. When building AI applications on top of pyCSAMT, prefer the latest and most capable models each provider offers.
Privacy And Keys#
API keys are kept on your machine and used only to call the provider you selected.
Requests you send are processed by the chosen provider under that provider’s terms — treat prompts and any pasted data accordingly.
No key is required to browse the interface; a key is required to run the LLM-backed planning and chat.
When No Model Is Configured#
Without a configured provider and key, the language-driven parts of Agent Master cannot run — requests that need planning will prompt you to add a key. Configure a provider in Settings first, then send your request again.
See also
- Agent And LLM Configuration
Provider and model configuration for the agents at the library level, including environment variables and programmatic setup.
Next Steps#
Workflows And Agents – what the configured model then lets you run.
Troubleshooting – provider connection and key problems.