Skip to main content

Models and providers

Pick the models NeoAgent thinks with — local, API-key, or account-backed.

NeoAgent supports local models, API-key providers, and several account-backed providers. Provider credentials stay on the server.

⚙️ Configure a model​

Open Settings > Models to choose chat and routing defaults. Add provider credentials on the admin dashboard's AI Providers page, with neoagent env, or during neoagent setup.

Desktop installs

Installs made from the desktop app have no neoagent command on PATH: the runtime lives under ~/.neoagent and is managed from Settings > Server. Use the admin dashboard for provider credentials there.

🏠 Local models​

Ollama does not require a hosted-model API key. Set its server URL with OLLAMA_URL, then select an available Ollama model in NeoAgent. The Ollama process may run on the NeoAgent host or another reachable machine.

🔑 API-key providers​

NeoAgent includes providers for Anthropic, OpenAI, Google Gemini, xAI, MiniMax, NVIDIA NIM, OpenRouter, and OpenAI-compatible endpoints. Available models are loaded through the configured provider rather than maintained as a fixed list in the documentation.

Configure a custom endpoint with both OPENAI_COMPATIBLE_BASE_URL and OPENAI_COMPATIBLE_API_KEY, either in the environment or on the admin dashboard's AI Providers page. The endpoint must implement the OpenAI Chat Completions and model-listing APIs. Putting that same token in OPENAI_API_KEY does not enable official OpenAI (api.openai.com).

🪪 Account-backed providers​

The CLI can authenticate supported developer subscriptions:

neoagent login github-copilot
neoagent login openai-codex
neoagent login claude-code
neoagent login grok-oauth

These login flows are separate from ordinary API-key providers.

🎯 Model assignment​

The default model is selected in settings. Individual agents and scheduled tasks can override it. A model used for tools must support the tool-calling contract expected by its NeoAgent provider.

Some features use separate providers:

  • Embeddings for memory search use a configured supported embedding provider and fall back to lexical retrieval when embeddings are unavailable.
  • Voice transcription can use Deepgram when enabled.
  • Image generation and analysis depend on the selected provider and model.

🔐 Credential handling​

API keys and account tokens are stored under the NeoAgent runtime directory on the server. Do not put credentials in task prompts, skills, screenshots, issue reports, or chat messages.

Use the environment CLI when terminal administration is more appropriate:

neoagent env list

env list masks secrets. The env get, env set, and env unset commands accept a variable name; env get prints the selected value, so avoid running it in recorded terminals or shared shells.

🔗 Related​