Models & Providers

OpenAI

Connect hosted OpenAI models for planning, supervision, and direct chat.

Overview

Providers decide where planning, supervision, and direct chat actually run. In practice, most local-first setups start with Ollama for local work and add hosted providers only where stronger reasoning is worth it.

Operator Note

Overseer works best when each layer stays explicit: the orchestrator routes, supervisors execute, tools return evidence, and traces show what actually happened.

Getting Started

Start by making one provider reliable. Confirm the runtime is reachable, then map it to the agents or supervisors that should use it.

Configuration

Document model names, timeouts, and fallback expectations before you spread a provider across multiple agents.

Examples

provider.toml

name = "ollama"
base_url = "http://localhost:11434"
default_model = "qwen2.5:7b-instruct"
role = "local-orchestration"

Next Steps

Continue with Desktop Setup or head back to the documentation home.