Choose a provider and model

Choose interactively

Open /provider to switch among local OpenAI-compatible servers, OpenAI API, OpenRouter, Vercel AI Gateway, CheaperInference, Anthropic, and OpenAI Codex, then /models [filter] to browse that endpoint’s catalog. /settings brings provider, model, theme, and API-key settings together.

› /provider
› /models coding
› /settings

Supported providers

Provider	Credential
`local`	None. Any OpenAI-compatible server; defaults to Ollama at `http://localhost:11434/v1`.
`openai`	`OPENAI_API_KEY`
`openai-codex`	ChatGPT subscription through device login.
`anthropic`	`ANTHROPIC_API_KEY`. Uses the native Messages API.
`openrouter`	`OPENROUTER_API_KEY`
`vercel`	`AI_GATEWAY_API_KEY` (Vercel AI Gateway)
`cheaperinference`	`CHEAPERINFERENCE_API_KEY`

Configure from the shell

Setting	Flag / environment
Model	`--model NAME` / `ORCA_MODEL`
Endpoint	`--base-url URL` / `ORCA_BASE_URL`
API key	`--api-key KEY` / selected provider’s environment key: OpenAI `OPENAI_API_KEY`, OpenRouter `OPENROUTER_API_KEY`, Vercel `AI_GATEWAY_API_KEY`, CheaperInference `CHEAPERINFERENCE_API_KEY`, or Anthropic `ANTHROPIC_API_KEY`
OpenRouter	`--openrouter` / `OPENROUTER_API_KEY`
Anthropic	`--anthropic` / `ANTHROPIC_API_KEY`
Web tools	`--firecrawl-key KEY` / `FIRECRAWL_API_KEY`

Flags and environment values take precedence over saved values. An explicit --api-key takes precedence over the selected provider’s environment key, which in turn takes precedence over its saved key.

Saved configuration

Orcacode stores the provider, last model per provider, theme, API keys, extension settings, MCP servers, skills overrides, and workspace-scoped approvals in config.json. The default directory is ~/.config/orcacode on Unix-like systems, $XDG_CONFIG_HOME/orcacode when set, or %APPDATA%\orcacode on Windows. Set ORCA_CONFIG_DIR to override the directory. For an OpenAI ChatGPT subscription, select openai-codex in /provider; Orcacode opens a device login flow and can import an existing official Codex login as a fallback.

Start here

Getting started

Work safely

Configure

Use the agent

Extend

Embed

Automate

Under the hood

Reference

Choose a provider and model

Choose interactively

Supported providers

Configure from the shell

Saved configuration

​Choose interactively

​Supported providers

​Configure from the shell

​Saved configuration

Choose interactively

Supported providers

Configure from the shell

Saved configuration