secrets.example.sh

#!/bin/bash

# Copy this file to ~/.ai-runner/secrets.sh and fill in your values.
# (Legacy: ~/.claude-switcher/secrets.sh is also supported for backward compatibility)

# AWS Bedrock Credentials
# See: https://code.claude.com/docs/en/amazon-bedrock
# Multiple authentication methods supported:

# Option 1: AWS Bedrock API Key (recommended for simplicity)
# export AWS_BEARER_TOKEN_BEDROCK="your-bedrock-api-key"

# Option 2: AWS Access Keys
# export AWS_ACCESS_KEY_ID="your_access_key"
# export AWS_SECRET_ACCESS_KEY="your_secret_key"
# export AWS_SESSION_TOKEN="your_session_token"  # Optional, for temporary credentials

# Option 3: AWS Profile
# export AWS_PROFILE="your-profile-name"

# Required for all AWS auth methods:
# export AWS_REGION="us-west-2"

# Google Vertex AI Credentials
# See: https://code.claude.com/docs/en/google-vertex-ai
# export ANTHROPIC_VERTEX_PROJECT_ID="your_gcp_project_id"
# export CLOUD_ML_REGION="global"  # or "us-east5", "us-central1", etc.

# Google Cloud Authentication Methods (in precedence order):
# 
# Method 1: Service Account Key File (highest precedence)
#   Recommended for production/CI environments
#   export GOOGLE_APPLICATION_CREDENTIALS="/path/to/service-account-key.json"
#
# Method 2: Application Default Credentials
#   Recommended for local development
#   Run: gcloud auth application-default login
#   No additional environment variable needed
#
# Method 3: gcloud User Credentials (lowest precedence)
#   Fallback for basic authentication
#   Run: gcloud auth login
#   No additional environment variable needed
#
# AI Runner automatically detects and uses the appropriate method
# based on what's available in your environment.

# Optional: Set specific regions for each model (defaults to CLOUD_ML_REGION if not set)
# export VERTEX_REGION_CLAUDE_3_5_SONNET="us-east5"
# export VERTEX_REGION_CLAUDE_3_5_HAIKU="us-east5"
# export VERTEX_REGION_CLAUDE_3_7_SONNET="us-east5"
# export VERTEX_REGION_CLAUDE_4_0_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_4_0_SONNET="us-east5"
# export VERTEX_REGION_CLAUDE_4_1_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_4_5_SONNET="us-east5"
# export VERTEX_REGION_CLAUDE_4_5_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_4_6_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_4_7_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_4_8_OPUS="europe-west1"
# export VERTEX_REGION_CLAUDE_5_OPUS="europe-west1"

# Anthropic API Key
# See: https://console.anthropic.com/
# export ANTHROPIC_API_KEY="sk-ant-..."

# Vercel AI Gateway Credentials
# See: https://vercel.com/ai-gateway
# Route Claude Code through Vercel AI Gateway for failover and unified billing
# Get your token from Vercel dashboard: https://vercel.com/dashboard/~/ai
# export VERCEL_AI_GATEWAY_TOKEN="vck_..."
# export VERCEL_AI_GATEWAY_URL="https://ai-gateway.vercel.sh"  # Default, can be customized

# Microsoft Foundry on Azure Credentials
# See: https://code.claude.com/docs/en/microsoft-foundry
# Announced Nov 18, 2024: https://www.anthropic.com/news/claude-in-microsoft-foundry

# Option 1: API Key authentication
# export ANTHROPIC_FOUNDRY_API_KEY="your-azure-api-key"

# Option 2: Use Azure default credential chain (az login)
# If ANTHROPIC_FOUNDRY_API_KEY is not set, Azure default credentials will be used

# Required: Azure resource name or full base URL
# export ANTHROPIC_FOUNDRY_RESOURCE="your-resource-name"
# Or provide the full URL:
# export ANTHROPIC_FOUNDRY_BASE_URL="https://your-resource-name.services.ai.azure.com"

# ============================================================================
# Model Overrides (Optional)
# ============================================================================
# You can override the default models defined in config/models.sh here.
# These provider-specific variables are used by our scripts to set ANTHROPIC_MODEL
# at runtime based on which provider you're using.

# Fable 5 (--fable/--best) is the top tier, above Opus, on Anthropic providers only.
# Defaults are set in config/models.sh; uncomment below to override.

# AWS Bedrock Models
# export CLAUDE_MODEL_SONNET_AWS="global.anthropic.claude-sonnet-5"
# export CLAUDE_MODEL_OPUS_AWS="global.anthropic.claude-opus-5"
# export CLAUDE_MODEL_HAIKU_AWS="us.anthropic.claude-haiku-4-5-20251001-v1:0"
# export CLAUDE_MODEL_FABLE_AWS="global.anthropic.claude-fable-5"

# Google Vertex Models
# export CLAUDE_MODEL_SONNET_VERTEX="claude-sonnet-5"
# export CLAUDE_MODEL_OPUS_VERTEX="claude-opus-5"
# export CLAUDE_MODEL_HAIKU_VERTEX="claude-haiku-4-5@20251001"
# export CLAUDE_MODEL_FABLE_VERTEX="claude-fable-5"

# Anthropic API Models (also used by --pro)
# export CLAUDE_MODEL_SONNET_ANTHROPIC="claude-sonnet-5"
# export CLAUDE_MODEL_OPUS_ANTHROPIC="claude-opus-5"
# export CLAUDE_MODEL_HAIKU_ANTHROPIC="claude-haiku-4-5"
# export CLAUDE_MODEL_FABLE_ANTHROPIC="claude-fable-5"

# Microsoft Foundry/Azure Models (deployment names - must match your Azure deployments)
# export CLAUDE_MODEL_SONNET_AZURE="claude-sonnet-5"
# export CLAUDE_MODEL_OPUS_AZURE="claude-opus-5"
# export CLAUDE_MODEL_HAIKU_AZURE="claude-haiku-4-5"
# export CLAUDE_MODEL_FABLE_AZURE="claude-fable-5"

# Vercel AI Gateway Models (supports any model via Vercel AI Gateway)
# export CLAUDE_MODEL_SONNET_VERCEL="anthropic/claude-sonnet-5"
# export CLAUDE_MODEL_OPUS_VERCEL="anthropic/claude-opus-5"
# export CLAUDE_MODEL_HAIKU_VERCEL="anthropic/claude-haiku-4.5"
# export CLAUDE_MODEL_FABLE_VERCEL="anthropic/claude-fable-5"
# Non-Anthropic models (see https://vercel.com/ai-gateway/models):
# export CLAUDE_MODEL_SONNET_VERCEL="xai/grok-code-fast-1"
# export CLAUDE_MODEL_SONNET_VERCEL="openai/gpt-5.2-codex"

# ============================================================================
# Small/Fast Model Overrides (Optional)
# ============================================================================
# Claude Code uses a "small/fast model" for background operations like file ops
# and sub-agents. By default, this uses each provider's Haiku model.
# You can override these if you want to use a different model for background tasks.

# AWS Bedrock Small/Fast Model
# export CLAUDE_SMALL_FAST_MODEL_AWS="us.anthropic.claude-haiku-4-5-20251001-v1:0"

# Google Vertex Small/Fast Model
# export CLAUDE_SMALL_FAST_MODEL_VERTEX="claude-haiku-4-5@20251001"

# Anthropic API Small/Fast Model
# export CLAUDE_SMALL_FAST_MODEL_ANTHROPIC="claude-haiku-4-5"

# Microsoft Foundry/Azure Small/Fast Model (deployment name)
# export CLAUDE_SMALL_FAST_MODEL_AZURE="claude-haiku-4-5"

# Vercel AI Gateway Small/Fast Model
# export CLAUDE_SMALL_FAST_MODEL_VERCEL="anthropic/claude-haiku-4.5"
# For non-Anthropic models, set to match your main model to avoid provider mixing:
# export CLAUDE_SMALL_FAST_MODEL_VERCEL="xai/grok-code-fast-1"

# ============================================================================
# Ollama Configuration (Local, Free)
# ============================================================================
# Ollama provides free local AI with Anthropic API compatibility.
# See: https://docs.ollama.com/integrations/claude-code

# Ollama server URL (default: http://localhost:11434)
# export OLLAMA_HOST="http://localhost:11434"

# Ollama Model Tier Mappings
# Override the default models used for each tier (--high, --mid, --low)
# Recommended: Use models with 64K+ context for Claude Code compatibility
# export OLLAMA_MODEL_HIGH="qwen3:72b"           # For --opus/--high
# export OLLAMA_MODEL_MID="qwen3-coder:32b"      # For --sonnet/--mid
# export OLLAMA_MODEL_LOW="qwen3:8b"             # For --haiku/--low
#
# Cloud models (no GPU required, 198K context):
# export OLLAMA_MODEL_HIGH="minimax-m2.5:cloud"  # 80% SWE-bench, best coding
# export OLLAMA_MODEL_MID="glm-5:cloud"          # 78% SWE-bench, strong reasoning

# ============================================================================
# LM Studio Configuration (Local)
# ============================================================================
# LM Studio runs local models with Anthropic API compatibility.
# Especially fast on Apple Silicon with MLX models.
# See: https://lmstudio.ai/blog/claudecode

# LM Studio server URL (default: http://localhost:1234)
# export LMSTUDIO_HOST="http://localhost:1234"

# LM Studio Model Tier Mappings
# By default, AI Runner uses the first loaded model for all tiers.
# Override to use different models per tier:
# export LMSTUDIO_MODEL_HIGH="openai/gpt-oss-20b"
# export LMSTUDIO_MODEL_MID="openai/gpt-oss-20b"
# export LMSTUDIO_MODEL_LOW="ibm/granite-4-micro"

# ============================================================================
# OpenRouter Configuration
# ============================================================================
# OpenRouter provides access to 500+ models with a single API key.
# See: https://openrouter.ai

# export OPENROUTER_API_KEY="sk-or-..."

# OpenRouter Model Tier Mappings
# export ROUTER_MODEL_HIGH="anthropic/claude-opus-5"
# export ROUTER_MODEL_MID="anthropic/claude-sonnet-5"
# export ROUTER_MODEL_LOW="anthropic/claude-haiku-4.5"

# ============================================================================
# Codex CLI (OpenAI) Configuration
# ============================================================================
# Required for: ai --codex
# See: https://developers.openai.com/codex/cli

# OpenAI API Key (used by Codex CLI for default cloud access)
# export OPENAI_API_KEY="sk-..."
# export CODEX_API_KEY="..."       # Alternative to OPENAI_API_KEY

# Codex Model Tier Overrides
# export CODEX_MODEL_HIGH="gpt-5.4"
# export CODEX_MODEL_MID="gpt-5.3-codex"
# export CODEX_MODEL_LOW="gpt-5.4-mini"

# For custom providers (Azure OpenAI, OpenRouter, Mistral, etc.),
# configure ~/.codex/config.toml directly:
#   [model_providers.azure]
#   name = "Azure OpenAI"
#   base_url = "https://YOUR_RESOURCE.openai.azure.com/openai/v1"
#   env_key = "AZURE_OPENAI_API_KEY"
#
# Then use: ai --codex --azure  or  ai --codex --profile azure

# ============================================================================
# AI Runner Defaults
# ============================================================================

# Default provider (used when no provider flag specified)
# Options: ollama, pro, apikey, aws, vertex, azure, vercel, lmstudio
# export DEFAULT_PROVIDER="ollama"

# Tip: You can also save defaults interactively:
#   ai --aws --opus --set-default    # Save preferred provider+model
#   ai --clear-default               # Remove saved default

# Skip config migration prompt
# export AI_RUNNER_SKIP_MIGRATION=1

# Disable update checking
# export AI_NO_UPDATE_CHECK=1

# Custom config directory
# export AI_RUNNER_CONFIG_DIR="$HOME/.ai-runner"