config.example.toml

# Claudex Configuration
# Copy this file to ~/.config/claudex/config.toml and edit

# Path to claude binary (default: "claude" from PATH)
# claude_binary = "/usr/local/bin/claude"

# Proxy settings
proxy_port = 13456
proxy_host = "127.0.0.1"

# Log level: trace, debug, info, warn, error
log_level = "info"

# Terminal hyperlinks (OSC 8): "auto" | true | false
# "auto" detects terminal support; true/false force on/off
hyperlinks = "auto"

# Model aliases (shorthand → full model name)
[model_aliases]
grok3 = "grok-3-beta"
gpt4o = "gpt-4o"
ds3 = "deepseek-chat"
claude = "claude-sonnet-4-20250514"

# ─── Profile Advanced Options Reference ────────────────
#
# All profile fields (most are optional with sensible defaults):
#
#   name            = "profile-name"           # (required) unique identifier
#   provider_type   = "OpenAICompatible"       # DirectAnthropic | OpenAICompatible | OpenAIResponses
#   base_url        = "https://..."            # (required) API endpoint
#   api_key         = "sk-..."                 # API key (or use api_key_keyring)
#   api_key_keyring = "keyring-entry"          # load api_key from system keyring
#   default_model   = "model-name"             # (required) default model for this profile
#   enabled         = true                     # enable/disable profile (default: true)
#   priority        = 100                      # fallback priority, higher = preferred (default: 100)
#   backup_providers = ["other-profile"]       # fallback profile names on failure
#   auth_type       = "api_key"                # "api_key" (default) | "oauth"
#   oauth_provider  = "openai"                 # required when auth_type = "oauth"
#   max_tokens      = 16384                    # cap max output tokens sent to provider (optional)
#   strip_params    = "auto"                   # "auto" | "none" | ["temperature", "top_p"] (default: "auto")
#
#   [profiles.models]                          # model slot mapping for Claude Code /model command
#   haiku  = "fast-model"                      # maps Claude haiku slot to this model
#   sonnet = "balanced-model"                  # maps Claude sonnet slot
#   opus   = "powerful-model"                  # maps Claude opus slot
#
#   [profiles.custom_headers]                  # extra HTTP headers per request
#   X-Custom-Header = "value"
#
#   [profiles.extra_env]                       # extra env vars passed to claude process
#   SOME_VAR = "value"
#
#   [profiles.query_params]                    # URL query params (e.g. Azure api-version)
#   api-version = "2024-12-01-preview"

# ─── Profiles ───────────────────────────────────────────

# Anthropic (native, no translation needed)
[[profiles]]
name = "anthropic"
provider_type = "DirectAnthropic"
base_url = "https://api.anthropic.com"
api_key = "YOUR_ANTHROPIC_KEY"   # or use api_key_keyring = "anthropic"
default_model = "claude-sonnet-4-20250514"
enabled = true
priority = 100

# MiniMax (Anthropic-compatible)
[[profiles]]
name = "minimax"
provider_type = "DirectAnthropic"
base_url = "https://api.minimax.io/anthropic"
api_key = "YOUR_API_KEY"
default_model = "claude-sonnet-4-20250514"
backup_providers = ["anthropic"]
enabled = true
priority = 90

# OpenRouter (needs translation — NOT Anthropic-compatible for passthrough)
[[profiles]]
name = "openrouter"
provider_type = "OpenAICompatible"
base_url = "https://openrouter.ai/api/v1"
api_key = "YOUR_OPENROUTER_KEY"
default_model = "anthropic/claude-sonnet-4"
enabled = true
priority = 80

# Grok (xAI, needs translation)
[[profiles]]
name = "grok"
provider_type = "OpenAICompatible"
base_url = "https://api.x.ai/v1"
api_key = "YOUR_XAI_KEY"
default_model = "grok-3-beta"
backup_providers = ["deepseek"]
enabled = true
priority = 100

# OpenAI (needs translation)
[[profiles]]
name = "chatgpt"
provider_type = "OpenAICompatible"
base_url = "https://api.openai.com/v1"
api_key = "YOUR_API_KEY"
default_model = "gpt-4o"
enabled = true
priority = 90

# DeepSeek (needs translation)
[[profiles]]
name = "deepseek"
provider_type = "OpenAICompatible"
base_url = "https://api.deepseek.com"
api_key = "YOUR_API_KEY"
default_model = "deepseek-chat"
backup_providers = ["grok"]
enabled = true
priority = 100

# Kimi / Moonshot (needs translation)
[[profiles]]
name = "kimi"
provider_type = "OpenAICompatible"
base_url = "https://api.moonshot.ai/v1"
api_key = "YOUR_API_KEY"
default_model = "kimi-k2-0905-preview"
enabled = true
priority = 70

# GLM / 智谱 (needs translation)
[[profiles]]
name = "glm"
provider_type = "OpenAICompatible"
base_url = "https://api.z.ai/api/paas/v4"
api_key = "YOUR_API_KEY"
default_model = "glm-4.6"
enabled = true
priority = 70

# ─── OAuth Subscription Profiles ───────────────────────
# Login first: claudex auth login <provider> --profile <name>

# Claude Max (via Claude Code's own OAuth)
[[profiles]]
name = "claude-max"
provider_type = "DirectAnthropic"
base_url = "https://api.claude.ai"
default_model = "claude-sonnet-4-20250514"
auth_type = "oauth"
oauth_provider = "claude"
enabled = false

# ChatGPT Plus/Pro / Codex subscription (via OpenAI OAuth → Responses API)
# Supported models: gpt-5.3-codex, gpt-5.1-codex, codex-mini-latest
# Note: gpt-4o is NOT supported on the Codex ChatGPT endpoint
# strip_params = "auto" will auto-detect chatgpt.com and strip temperature/top_p/top_k
[[profiles]]
name = "codex-sub"
provider_type = "OpenAIResponses"
base_url = "https://chatgpt.com/backend-api/codex"
default_model = "gpt-5.3-codex"
auth_type = "oauth"
oauth_provider = "openai"
enabled = false
# ChatGPT-Account-ID is auto-extracted from Codex CLI's auth.json
# Or set manually: [profiles.custom_headers]
# ChatGPT-Account-ID = "your-account-id"

# Gemini Pro (via Google OAuth)
[[profiles]]
name = "gemini-sub"
provider_type = "OpenAICompatible"
base_url = "https://generativelanguage.googleapis.com/v1beta/openai"
default_model = "gemini-2.5-pro"
auth_type = "oauth"
oauth_provider = "google"
enabled = false

# GitHub Copilot (via GitHub Device Code Flow)
[[profiles]]
name = "copilot"
provider_type = "OpenAICompatible"
base_url = "https://api.githubcopilot.com"
default_model = "gpt-4o"
auth_type = "oauth"
oauth_provider = "github"
enabled = false

# GitLab Duo (via GITLAB_TOKEN env)
# Login: claudex auth login gitlab --profile gitlab-duo
# Requires: export GITLAB_TOKEN=glpat-...
[[profiles]]
name = "gitlab-duo"
provider_type = "OpenAICompatible"
base_url = "https://gitlab.com/api/v4/ai/llm/proxy"
default_model = "claude-sonnet-4-20250514"
auth_type = "oauth"
oauth_provider = "gitlab"
enabled = false

# ─── Cloud Providers (API Key) ────────────────────────

# Azure OpenAI (uses api-key header + query_params for api-version)
# Auto-detected by base_url containing "openai.azure.com"
[[profiles]]
name = "azure-openai"
provider_type = "OpenAICompatible"
base_url = "https://YOUR_RESOURCE.openai.azure.com/openai/deployments/YOUR_DEPLOYMENT"
api_key = "YOUR_AZURE_KEY"
default_model = "gpt-4o"
enabled = false
priority = 80
[profiles.query_params]
api-version = "2024-12-01-preview"

# Google Vertex AI (DirectAnthropic endpoint, use gcloud auth print-access-token)
# base_url format: https://REGION-aiplatform.googleapis.com/v1/projects/PROJECT/locations/REGION/publishers/anthropic/models
# Refresh token: gcloud auth print-access-token
[[profiles]]
name = "vertex-ai"
provider_type = "DirectAnthropic"
base_url = "https://us-east5-aiplatform.googleapis.com/v1/projects/YOUR_PROJECT/locations/us-east5/publishers/anthropic/models"
api_key = "YOUR_GCLOUD_TOKEN"
default_model = "claude-sonnet-4@20250514"
enabled = false
priority = 80

# Groq (fast inference)
[[profiles]]
name = "groq"
provider_type = "OpenAICompatible"
base_url = "https://api.groq.com/openai/v1"
api_key = "YOUR_GROQ_KEY"
default_model = "llama-3.3-70b-versatile"
enabled = false
priority = 70

# Mistral AI
[[profiles]]
name = "mistral"
provider_type = "OpenAICompatible"
base_url = "https://api.mistral.ai/v1"
api_key = "YOUR_MISTRAL_KEY"
default_model = "mistral-large-latest"
enabled = false
priority = 70

# Together AI
[[profiles]]
name = "together"
provider_type = "OpenAICompatible"
base_url = "https://api.together.xyz/v1"
api_key = "YOUR_TOGETHER_KEY"
default_model = "meta-llama/Llama-3.3-70B-Instruct-Turbo"
enabled = false
priority = 70

# Perplexity (online search + LLM)
[[profiles]]
name = "perplexity"
provider_type = "OpenAICompatible"
base_url = "https://api.perplexity.ai"
api_key = "YOUR_PPLX_KEY"
default_model = "sonar-pro"
enabled = false
priority = 60

# Cerebras (fast inference)
[[profiles]]
name = "cerebras"
provider_type = "OpenAICompatible"
base_url = "https://api.cerebras.ai/v1"
api_key = "YOUR_CEREBRAS_KEY"
default_model = "llama-3.3-70b"
enabled = false
priority = 70

# ─── Local Models (via Ollama / vLLM / LM Studio) ──────

# Ollama
[[profiles]]
name = "local-qwen"
provider_type = "OpenAICompatible"
base_url = "http://localhost:11434/v1"
api_key = ""
default_model = "qwen2.5:72b"
enabled = false
priority = 50

# LM Studio
[[profiles]]
name = "lm-studio"
provider_type = "OpenAICompatible"
base_url = "http://localhost:1234/v1"
api_key = "lm-studio"
default_model = "local-model"
enabled = false
priority = 50

# vLLM / llama.cpp server
[[profiles]]
name = "local-llama"
provider_type = "OpenAICompatible"
base_url = "http://localhost:8000/v1"
api_key = ""
default_model = "llama-3.3-70b"
enabled = false
priority = 50

# Bedrock via LiteLLM proxy
# Run: litellm --model bedrock/anthropic.claude-sonnet-4-20250514-v2:0
# See: https://docs.litellm.ai/docs/providers/bedrock
# [[profiles]]
# name = "bedrock"
# provider_type = "OpenAICompatible"
# base_url = "http://localhost:4000/v1"
# api_key = "sk-litellm"
# default_model = "bedrock/anthropic.claude-sonnet-4-20250514-v2:0"
# enabled = false

# ─── Smart Router (optional) ───────────────────────────

[router]
enabled = false
profile = "local-qwen"              # reuse a profile's base_url + api_key for classification
model = "qwen2.5:3b"                # override model (optional, defaults to profile's default_model)

[router.rules]
code = "deepseek"
analysis = "grok"
creative = "chatgpt"
search = "kimi"
math = "deepseek"
default = "grok"

# ─── Context Engine (optional) ─────────────────────────

[context.compression]
enabled = false
threshold_tokens = 50000
keep_recent = 10
profile = "local-qwen"              # reuse a profile's base_url + api_key for summarization
model = "qwen2.5:3b"                # override model (optional)

[context.sharing]
enabled = false
max_context_size = 2000

[context.rag]
enabled = false
index_paths = ["./src", "./docs"]
profile = "local-qwen"              # reuse a profile's base_url + api_key for embeddings
model = "nomic-embed-text"           # embedding model
chunk_size = 512
top_k = 5