Model connections

SeedCode CLI providers: Your models.
Your choice.

Connect a built-in default, cloud provider, local Ollama model or any OpenAI-compatible endpoint.

ProviderApi keyDetails
DefaultNoBuilt-in connection; also supports OPENROUTER_API_KEY / SEEDCODE_DEFAULT_API_KEY.
OpenRouterYesUse your own OpenRouter key and model.
FreeModel ClaudeYesFreeModel backend, Claude models.
FreeModel CodexYesFreeModel backend, Codex models.
AeroLinkYesAeroLink backend.
OllamaNoRun local models. Seed Code starts ollama serve when needed.
CustomDependsAny OpenAI-compatible endpoint: name, base URL, key and model.

Local models with Ollama

Install Ollama, pull a model, then select it inside Seed Code.

ollama pull qwen3:4b
seedcode
/provider → Ollama
/model → qwen3:4b
A local Ollama session is never silently moved to a cloud provider. Local-to-cloud failover requires the explicit allow_local_to_cloud opt-in.

Execution profiles

Small

Compact context

Short sequential instructions, explicit tool protocol and tighter call/output budgets for smaller local models.

Standard

Balanced execution

General coding models with normal context, step and tool-call budgets.

Advanced

Full capability

Strong models receive fuller context and larger orchestration budgets.