| summary |
read_when |
| Index of AI provider docs (OpenAI, Anthropic, Gemini, MiniMax, Grok, Ollama). |
| choosing or configuring AI providers for Peekaboo |
| looking for provider-specific plans or status |
|
Providers index
- OpenAI —
openai.md: architecture, migration status, and guidance for adding models.
- Anthropic —
anthropic.md: Fable/Claude models, output limits, generation settings, and credentials.
- Google — configured with
GEMINI_API_KEY; supports Gemini 3.1 Pro Preview and Gemini 3 Flash.
- MiniMax — configured with
MINIMAX_API_KEY; supports MiniMax M2.7 through the Anthropic-compatible API.
- MiniMax China — use
minimax-cn/... with MINIMAX_CN_API_KEY or the shared MINIMAX_API_KEY; routes to api.minimaxi.com.
- Grok —
grok.md: Grok 4 implementation guide and checkpoints.
- Ollama —
ollama.md: local model configuration; ollama-models.md for model catalog notes.
Use docs/providers.md as the central reference for the user-facing provider list,
configuration syntax, and environment variable reference.
Capability quick-compare
| Provider |
Tools |
Vision |
Streaming |
Local/offline |
Auth |
| OpenAI |
Yes (function/tool calling) |
Yes |
Yes |
No |
API key or OAuth |
| Anthropic |
Yes |
Yes |
Model-dependent; Fable/Opus 4.8 currently non-streaming |
No |
API key or OAuth (Claude Pro/Max) |
| Google |
Yes |
Yes |
Yes |
No |
API key |
| MiniMax |
Yes |
No |
Yes |
No |
API key |
| MiniMax China |
Yes |
No |
Yes |
No |
API key |
| Grok |
Yes |
Limited |
Yes |
No |
API key |
| Ollama |
Yes (via local server) |
Model-dependent |
Yes |
Yes (local) |
None (local daemon) |
| LM Studio |
Yes (OpenAI-compatible local server) |
Model-dependent |
Yes |
Yes (local) |
None by default |
See individual pages for model lists, quirks, and test coverage expectations.
