@kairohq/ai

Provider-abstracted summarization and embedding helpers for Kairo.

Workspace AI settings live in .kairo/config.json under ai. kairo init writes Anthropic API-key config by default, --ai-provider none disables AI, and KAIRO_AI_PROVIDER / KAIRO_AI_MODEL / KAIRO_AI_EMBEDDING_MODEL / KAIRO_AI_API_KEY_ENV / KAIRO_AI_BASE_URL can override runtime selection.

{
  "ai": {
    "provider": "anthropic",
    "model": "claude-sonnet-4-5",
    "apiKeyEnv": "ANTHROPIC_API_KEY",
    "baseUrl": "https://api.anthropic.com",
    "authMode": "api-key"
  }
}

Provider setup order

Most common choices are listed first; the rest are alphabetical.

| Provider | Env var | Base URL | Notes | | --- | --- | --- | --- | | OpenAI | OPENAI_API_KEY | https://api.openai.com/v1 | API key via bearer auth. | | Anthropic | ANTHROPIC_API_KEY | https://api.anthropic.com | API key via x-api-key; authMode: "headless" reserves Workload Identity Federation config. | | Gemini | GEMINI_API_KEY | https://generativelanguage.googleapis.com | API key for Gemini API; Vertex ADC is future config work. | | OpenRouter | OPENROUTER_API_KEY | https://openrouter.ai/api/v1 | OpenAI-compatible chat and embedding surface; authMode: "headless" reserves OAuth config. | | Ollama | none | http://localhost:11434 | Local-first default candidate; no API key for local daemon. | | Amazon Bedrock | BEDROCK_API_KEY | region-specific https://bedrock-mantle.<region>.api.aws/v1 | OpenAI-compatible API-key path; authMode: "headless" reserves IAM/temporary credential config. | | Azure OpenAI | AZURE_OPENAI_API_KEY | resource-specific | Planned. authMode: "headless" reserves Microsoft Entra ID config. | | Cerebras | CEREBRAS_API_KEY | https://api.cerebras.ai/v1 | OpenAI-compatible chat surface. | | Cohere | COHERE_API_KEY | https://api.cohere.com | Planned native provider. | | Custom OpenAI-compatible | user-defined | user-defined | Set provider: "custom" and provide baseUrl. | | DeepSeek | DEEPSEEK_API_KEY | https://api.deepseek.com | OpenAI-compatible chat surface. | | Fireworks AI | FIREWORKS_API_KEY | https://api.fireworks.ai/inference/v1 | OpenAI-compatible chat surface. | | Groq | GROQ_API_KEY | https://api.groq.com/openai/v1 | OpenAI-compatible chat surface. | | Kilo Gateway | KILO_API_KEY | https://api.kilo.ai/api/gateway | Gateway for Kimi, MiniMax, DeepSeek, Grok, Qwen, Claude, GPT, Gemini, and more. | | LM Studio | LM_STUDIO_API_KEY | http://localhost:1234/v1 | Local OpenAI-compatible server; API key is optional. | | MiniMax | MINIMAX_API_KEY | https://api.minimax.io/v1 | OpenAI-compatible MiniMax models. | | Mistral | MISTRAL_API_KEY | https://api.mistral.ai/v1 | OpenAI-compatible chat and embedding surface. | | Moonshot Kimi | MOONSHOT_API_KEY | https://api.moonshot.ai/v1 | OpenAI-compatible Kimi/Kimi K2 models. | | Perplexity | PERPLEXITY_API_KEY | https://api.perplexity.ai | OpenAI-compatible chat surface with search-oriented models. | | Together AI | TOGETHER_API_KEY | https://api.together.xyz/v1 | OpenAI-compatible hosted open models. | | Vertex AI | ADC | region/project-specific | Planned headless auth via Google Application Default Credentials. | | xAI Grok | XAI_API_KEY | https://api.x.ai/v1 | OpenAI-compatible Grok models. |