Skip to content

Providers ​

This index lists every first-party provider Rho ships. Shared concepts such as credential stores, model selection, first-run setup, and environment overrides live in authentication and models. Each provider page covers login, logout, auth modes, and model notes for that backend.

ProviderAuth modesPage
openaiapi-keyOpenAI
openai-codexcodexOpenAI (Codex OAuth)
anthropicanthropic-api-keyAnthropic
googlegoogle-api-keyGoogle Gemini
github-copilotgithub-copilotGitHub Copilot
xaixai-api-key, xai-oauthSpaceXAI
poolsidepoolside-api-keyPoolside
openrouteropenrouter-api-key, openrouter-oauthOpenRouter
ollamanone, optional ollama-api-keyOllama
ollama-cloudollama-cloud-api-key, ollama-cloud-deviceOllama Cloud
moonshotmoonshot-api-keyMoonshot and Kimi Code
kimi-codekimi-oauthMoonshot and Kimi Code
qwen-token-planqwen-token-plan-api-keyQwen Token Plan
metameta-api-key, meta-museMeta Model API
minimaxminimax-api-keyMiniMax
mistralmistral-api-keyMistral AI
opencode-goopencode-go-api-keyOpenCode Go
opencodeopencode-api-keyOpenCode Zen
typesafetypesafe-api-keyTypeSafe

User-defined OpenAI-compatible hosts can be created from /login (Custom · Chat Completions or Custom · Responses) or by adding [providers.custom.<name>] with a base_url. They default to Chat Completions; set api = "responses" in config, or choose Custom · Responses in /login, to use the Responses API. An API key is optional. See Custom OpenAI-compatible hosts.

Large inference requests are automatically gzipped at verified endpoints, and custom hosts can opt in to zstd or gzip with request_compression. See request compression for eligibility and the provider support audit.

  1. Read authentication and models for credential storage and model selection.
  2. Open the provider page that matches your account.
  3. Run rho and finish setup with /login and /model, or pass provider flags to automation.