providers.yaml

# Tokenomics Provider Configuration
#
# Defines known upstream AI providers. Policies reference these by name
# in their "providers" map, inheriting connection details automatically.
#
# Each provider specifies:
#   upstream_url  - Base URL for the provider's API
#   api_key_env   - Environment variable holding the API key
#   base_url_env  - Environment variable for client base URL (used by init/run)
#   auth_scheme   - How to send the key: "bearer" (default), "header", "query"
#   auth_header   - Custom header name (default: "Authorization")
#   headers       - Extra headers sent with every request
#   models        - Known model prefixes (informational, for documentation)
#   chat_path     - Override path for chat completions endpoint

providers:

  # ── OpenAI ─────────────────────────────────────────────────────────────
  openai:
    upstream_url: https://api.openai.com
    api_key_env: OPENAI_API_KEY
    base_url_env: OPENAI_BASE_URL
    models:
      - gpt-4o
      - gpt-4o-mini
      - gpt-4-turbo
      - gpt-4
      - gpt-3.5-turbo
      - o1
      - o1-mini
      - o1-pro
      - o3
      - o3-mini
      - o4-mini

  # ── Anthropic ──────────────────────────────────────────────────────────
  anthropic:
    upstream_url: https://api.anthropic.com
    api_key_env: ANTHROPIC_API_KEY
    base_url_env: ANTHROPIC_BASE_URL
    auth_scheme: header
    auth_header: x-api-key
    headers:
      anthropic-version: "2023-06-01"
    chat_path: /v1/messages
    models:
      - claude-opus-4-6
      - claude-sonnet-4-6
      - claude-haiku-4-5
      - claude-3.5-sonnet
      - claude-3.5-haiku
      - claude-3-opus
      - claude-3-sonnet
      - claude-3-haiku

  # ── Azure OpenAI ───────────────────────────────────────────────────────
  # NOTE: Replace {resource} and {deployment} with your Azure values.
  # upstream_url format: https://{resource}.openai.azure.com/openai/deployments/{deployment}
  azure_openai:
    upstream_url: https://{resource}.openai.azure.com
    api_key_env: AZURE_OPENAI_API_KEY
    base_url_env: AZURE_OPENAI_ENDPOINT
    auth_scheme: header
    auth_header: api-key
    headers:
      api-version: "2024-10-21"
    models:
      - gpt-4o
      - gpt-4o-mini
      - gpt-4
      - gpt-35-turbo
      - o1
      - o1-mini

  # ── Google Gemini (AI Studio) ──────────────────────────────────────────
  google_gemini:
    upstream_url: https://generativelanguage.googleapis.com
    api_key_env: GOOGLE_API_KEY
    base_url_env: GEMINI_BASE_URL
    auth_scheme: query
    chat_path: /v1beta/models/{model}:generateContent
    models:
      - gemini-2.5-pro
      - gemini-2.5-flash
      - gemini-2.0-flash
      - gemini-2.0-flash-lite
      - gemini-1.5-pro
      - gemini-1.5-flash

  # ── Google Vertex AI ───────────────────────────────────────────────────
  # NOTE: Replace {project} and {location} with your GCP values.
  # Uses OAuth2 bearer token from gcloud auth.
  vertex_ai:
    upstream_url: https://{location}-aiplatform.googleapis.com
    api_key_env: VERTEX_AI_TOKEN
    base_url_env: VERTEX_AI_BASE_URL
    chat_path: /v1/projects/{project}/locations/{location}/publishers/google/models/{model}:generateContent
    models:
      - gemini-2.5-pro
      - gemini-2.5-flash
      - gemini-2.0-flash

  # ── Mistral AI ─────────────────────────────────────────────────────────
  mistral:
    upstream_url: https://api.mistral.ai
    api_key_env: MISTRAL_API_KEY
    base_url_env: MISTRAL_BASE_URL
    models:
      - mistral-large-latest
      - mistral-medium-latest
      - mistral-small-latest
      - open-mistral-nemo
      - open-mixtral-8x22b
      - codestral-latest
      - pixtral-large-latest

  # ── Cohere ─────────────────────────────────────────────────────────────
  cohere:
    upstream_url: https://api.cohere.com
    api_key_env: COHERE_API_KEY
    base_url_env: COHERE_BASE_URL
    chat_path: /v2/chat
    models:
      - command-r-plus
      - command-r
      - command-light
      - command-nightly

  # ── Groq ───────────────────────────────────────────────────────────────
  groq:
    upstream_url: https://api.groq.com/openai
    api_key_env: GROQ_API_KEY
    base_url_env: GROQ_BASE_URL
    models:
      - llama-3.3-70b-versatile
      - llama-3.1-8b-instant
      - llama-3.1-70b-versatile
      - mixtral-8x7b-32768
      - gemma2-9b-it

  # ── Together AI ────────────────────────────────────────────────────────
  together:
    upstream_url: https://api.together.xyz
    api_key_env: TOGETHER_API_KEY
    base_url_env: TOGETHER_BASE_URL
    models:
      - meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo
      - meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo
      - meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo
      - mistralai/Mixtral-8x22B-Instruct-v0.1
      - Qwen/Qwen2.5-72B-Instruct-Turbo

  # ── Fireworks AI ───────────────────────────────────────────────────────
  fireworks:
    upstream_url: https://api.fireworks.ai/inference
    api_key_env: FIREWORKS_API_KEY
    base_url_env: FIREWORKS_BASE_URL
    models:
      - accounts/fireworks/models/llama-v3p1-405b-instruct
      - accounts/fireworks/models/llama-v3p1-70b-instruct
      - accounts/fireworks/models/mixtral-8x22b-instruct
      - accounts/fireworks/models/qwen2p5-72b-instruct

  # ── Perplexity ─────────────────────────────────────────────────────────
  perplexity:
    upstream_url: https://api.perplexity.ai
    api_key_env: PERPLEXITY_API_KEY
    base_url_env: PERPLEXITY_BASE_URL
    models:
      - sonar-pro
      - sonar
      - sonar-reasoning-pro
      - sonar-reasoning

  # ── DeepSeek ───────────────────────────────────────────────────────────
  deepseek:
    upstream_url: https://api.deepseek.com
    api_key_env: DEEPSEEK_API_KEY
    base_url_env: DEEPSEEK_BASE_URL
    models:
      - deepseek-chat
      - deepseek-reasoner

  # ── xAI (Grok) ────────────────────────────────────────────────────────
  xai:
    upstream_url: https://api.x.ai
    api_key_env: XAI_API_KEY
    base_url_env: XAI_BASE_URL
    models:
      - grok-2
      - grok-2-mini
      - grok-3
      - grok-3-mini

  # ── OpenRouter ─────────────────────────────────────────────────────────
  # Routes to multiple providers behind a single API. Models use
  # provider/model format (e.g., "openai/gpt-4o", "anthropic/claude-3-opus").
  openrouter:
    upstream_url: https://openrouter.ai/api
    api_key_env: OPENROUTER_API_KEY
    base_url_env: OPENROUTER_BASE_URL
    models:
      - openai/gpt-4o
      - anthropic/claude-3-opus
      - meta-llama/llama-3.1-405b-instruct
      - google/gemini-pro-1.5

  # ── Ollama (local) ─────────────────────────────────────────────────────
  # No API key needed for local inference.
  ollama:
    upstream_url: http://localhost:11434
    api_key_env: ""
    base_url_env: OLLAMA_BASE_URL
    auth_scheme: header
    auth_header: ""
    chat_path: /api/chat
    models:
      - llama3.1
      - llama3.2
      - mistral
      - codellama
      - phi3
      - gemma2
      - qwen2.5

  # ── vLLM (self-hosted) ─────────────────────────────────────────────────
  # OpenAI-compatible API served by vLLM.
  vllm:
    upstream_url: http://localhost:8000
    api_key_env: VLLM_API_KEY
    base_url_env: VLLM_BASE_URL
    models: []

  # ── LiteLLM Proxy ──────────────────────────────────────────────────────
  # If running LiteLLM as a unified gateway.
  litellm:
    upstream_url: http://localhost:4000
    api_key_env: LITELLM_API_KEY
    base_url_env: LITELLM_BASE_URL
    models: []

  # ── AWS Bedrock (OpenAI-compatible gateways commonly front this) ─────────
  bedrock:
    upstream_url: https://bedrock-runtime.{region}.amazonaws.com
    api_key_env: AWS_BEARER_TOKEN_BEDROCK
    base_url_env: BEDROCK_BASE_URL
    models:
      - anthropic.claude-3-5-sonnet-20241022-v2:0
      - anthropic.claude-3-7-sonnet-20250219-v1:0
      - amazon.nova-pro-v1:0
      - meta.llama3-1-70b-instruct-v1:0

  # ── Replicate ─────────────────────────────────────────────────────────────
  replicate:
    upstream_url: https://api.replicate.com
    api_key_env: REPLICATE_API_TOKEN
    base_url_env: REPLICATE_BASE_URL
    models:
      - meta/meta-llama-3-70b-instruct
      - mistralai/mistral-7b-instruct-v0.2

  # ── NVIDIA NIM ────────────────────────────────────────────────────────────
  nvidia_nim:
    upstream_url: https://integrate.api.nvidia.com
    api_key_env: NVIDIA_API_KEY
    base_url_env: NVIDIA_NIM_BASE_URL
    models:
      - meta/llama-3.1-70b-instruct
      - nvidia/llama-3.1-nemotron-70b-instruct

  # ── SambaNova ─────────────────────────────────────────────────────────────
  sambanova:
    upstream_url: https://api.sambanova.ai/v1
    api_key_env: SAMBANOVA_API_KEY
    base_url_env: SAMBANOVA_BASE_URL
    models:
      - Meta-Llama-3.1-70B-Instruct
      - Qwen2.5-72B-Instruct

  # ── Cerebras ──────────────────────────────────────────────────────────────
  cerebras:
    upstream_url: https://api.cerebras.ai/v1
    api_key_env: CEREBRAS_API_KEY
    base_url_env: CEREBRAS_BASE_URL
    models:
      - llama-3.3-70b
      - qwen-2.5-72b

  # ── AI21 ──────────────────────────────────────────────────────────────────
  ai21:
    upstream_url: https://api.ai21.com/studio/v1
    api_key_env: AI21_API_KEY
    base_url_env: AI21_BASE_URL
    models:
      - jamba-1.5-large
      - jamba-1.5-mini

  # ── Moonshot (Kimi) ───────────────────────────────────────────────────────
  moonshot:
    upstream_url: https://api.moonshot.ai/v1
    api_key_env: MOONSHOT_API_KEY
    base_url_env: MOONSHOT_BASE_URL
    models:
      - moonshot-v1-8k
      - moonshot-v1-128k

  # ── Anyscale Endpoints ────────────────────────────────────────────────────
  anyscale:
    upstream_url: https://api.endpoints.anyscale.com/v1
    api_key_env: ANYSCALE_API_KEY
    base_url_env: ANYSCALE_BASE_URL
    models:
      - meta-llama/Llama-3.1-70B-Instruct
      - mistralai/Mixtral-8x7B-Instruct-v0.1

  # ── LocalAI ───────────────────────────────────────────────────────────────
  localai:
    upstream_url: http://localhost:8081
    api_key_env: LOCALAI_API_KEY
    base_url_env: LOCALAI_BASE_URL
    models: []

  # ── LM Studio Local Server ────────────────────────────────────────────────
  lmstudio:
    upstream_url: http://localhost:1234
    api_key_env: LMSTUDIO_API_KEY
    base_url_env: LMSTUDIO_BASE_URL
    models: []