Admin Dashboard: Models

The Models page lets you configure LLM providers and endpoints. Each model entry defines how the engine connects to an LLM backend — you can have multiple models from different providers and assign each agent its own model.

Supported providers

Provider	Description
ollama	Local model inference via Ollama. Free, private, no API key needed. Requires Ollama installed on the host.
openai_compatible	Any API that follows OpenAI chat completions format. Works with OpenAI, DeepInfra, Together, Groq, vLLM, LiteLLM.
anthropic	Native Anthropic API. Supports Claude models with automatic message format conversion.
azure_openai	Azure-hosted OpenAI models. Uses deployment-based URLs and api-key header auth. Requires api_version.
google	Google Gemini models via the generateContent API. Uses x-goog-api-key header.
openrouter	OpenRouter aggregator. Access 100+ models via a single API key. Base URL is preset automatically.
deepseek	DeepSeek models (e.g. DeepSeek V3.2). OpenAI-compatible with preset base URL.
mistral	Mistral AI models (e.g. Mistral Medium 3). OpenAI-compatible with preset base URL.
xai	xAI Grok models (e.g. Grok 4.1). OpenAI-compatible with preset base URL.
zai	Z.ai GLM models (e.g. GLM-5). OpenAI-compatible with preset base URL.

Adding a model

Click “Add Model” and fill in the fields:

• Display name — a human-readable name used in the agent configuration dropdown.
• Provider — select from the supported providers above.
• Model name — the model identifier as expected by the provider API (e.g., llama3.2, claude-sonnet-4-20250514). For Azure OpenAI, this is the deployment name.
• Base URL — custom endpoint URL. Required for Ollama and third-party providers. Leave empty for providers with preset URLs (OpenAI, Anthropic, OpenRouter, DeepSeek, Mistral, xAI, Z.ai). For Azure OpenAI, this is your Azure resource URL (e.g., https://my-company.openai.azure.com).
• API Key — provider API key. Not needed for Ollama. Use the ${VAR} syntax when configuring via YAML.
• API Version — (Azure OpenAI only) the API version string, e.g. 2024-10-21.

Model validation

After adding a model, the engine attempts a test connection to verify the endpoint is reachable and the API key is valid. If the connection fails, the model is saved but marked with a warning indicator in the list.

Configuration examples

# Ollama (local, no API key needed)
models:
  llama-local:
    provider: ollama
    model_name: llama3.2
    base_url: "http://localhost:11434/v1"

# OpenAI-compatible (DeepInfra)
models:
  qwen-3-32b:
    provider: openai_compatible
    model_name: Qwen/Qwen3-32B
    base_url: "https://api.deepinfra.com/v1/openai"
    api_key: ${DEEPINFRA_API_KEY}

# Anthropic
models:
  claude-sonnet-4:
    provider: anthropic
    model_name: claude-sonnet-4-20250514
    api_key: ${ANTHROPIC_API_KEY}

# Azure OpenAI (deployment-based)
models:
  gpt4-azure:
    provider: azure_openai
    base_url: "https://my-company.openai.azure.com"
    model_name: "gpt-4o-deploy"
    api_version: "2024-10-21"
    api_key: ${AZURE_OPENAI_KEY}

# Google Gemini
models:
  gemini-pro:
    provider: google
    model_name: "gemini-3.1-pro"
    api_key: ${GOOGLE_API_KEY}

# OpenRouter (base_url is preset automatically)
models:
  openrouter-claude:
    provider: openrouter
    model_name: "anthropic/claude-sonnet-4-20250514"
    api_key: ${OPENROUTER_API_KEY}

# DeepSeek (preset base_url)
models:
  deepseek-v3:
    provider: deepseek
    model_name: "deepseek-chat"
    api_key: ${DEEPSEEK_API_KEY}

# Mistral (preset base_url)
models:
  mistral-medium:
    provider: mistral
    model_name: "mistral-medium-3"
    api_key: ${MISTRAL_API_KEY}

# xAI (preset base_url)
models:
  grok:
    provider: xai
    model_name: "grok-4.1"
    api_key: ${XAI_API_KEY}

# Z.ai / GLM (preset base_url)
models:
  glm-5:
    provider: zai
    model_name: "glm-5"
    api_key: ${ZAI_API_KEY}

---

What’s next

• MCP Servers
• Configuration: Models