Routing Status

Monitor your AI routing configuration

Providers

Models

Status

Active

Router Configuration

Default Model -

Think Model -

Background Model -

WebSearch Model -

Server Info

Address -

API Endpoints

POST /v1/messages

Send messages with automatic model routing

POST /v1/messages/count_tokens

Count tokens for messages

GET /api/models-config

Get models configuration (name, provider mappings)

GET /api/providers

Get providers configuration (name, type, API keys)

GET /api/config

Get router configuration (default, think, background models)

GET /api/config/json

Get full configuration (JSON format)

POST /api/config/json

Update full configuration (JSON format)

POST /api/restart

Restart the server to apply configuration changes

Providers

Connect and manage AI providers

Anthropic

Active

anthropic-native

3 models • API key configured

🔌

No providers configured

Add a provider to use more AI models

Provider Add

Enter your API key to start using

Which provider do you want to add?

Anthropic

Claude models

z.ai

GLM models

ZenMux

Unified API gateway

Minimax

MiniMax models

Kimi For Coding

Premium membership for Kimi

OpenAI

GPT models

Google Gemini

Gemini models (AI Studio)

Vertex AI

Gemini, Claude, Llama on GCP

OpenRouter

500+ models

Groq

Ultra-fast ⚡

Together AI

Open source

Fireworks AI

Fast inference

Deepinfra

Cost-effective

Cerebras

Ultra-fast ⚡

Nebius

AI platform

Moonshot AI

Kimi models

NovitaAI

GPU cloud

Baseten

ML deployment

Choose a name for this provider

Name

A name to identify this provider

Choose authentication method

☁️

Google Cloud Vertex AI Configuration

Project ID

Your Google Cloud Project ID

Location

Region where Vertex AI is enabled

ℹ️ Authentication:
Vertex AI uses Application Default Credentials (ADC). Before starting the server, run:
gcloud auth application-default login

API Key

Use your API key for authentication

OAuth (Claude Pro/Max)

Free for Claude Max subscribers

API Key

Your API key will be stored securely

🔐

OAuth Authentication

Click the button below to authenticate with your Claude Pro/Max account.

✓ Authorization window opened

Log in to your Claude Pro/Max account
Click "Allow" to authorize
Copy the authorization code
Paste it in the field below

Authorization Code

OAuth token saved!

For Claude Pro/Max subscribers only. Requires active subscription.

Add Settings

Optional - only if needed

Custom Endpoint (Optional)

Only enter if you want to use a different URL than the default endpoint

Models

Improve stability with multiple providers

claude-sonnet-4-5

Priority 1 anthropic/claude-sonnet-4-5

Priority 2 openrouter/anthropic/claude-sonnet-4-5

🎯

No models configured

Add a model to make it available via API

Add Model

Configure multiple providers to increase reliability

Choose a name for this model

Model Name

This name will be used in API requests

Map providers to actual models

Tries providers in order, automatically fails over to the next one Auto-switching enabled

Priority 1 Primary

Select Provider

Model Name

Actual model name used by the provider Model ID

Router Configuration

Configure different models for different scenarios

Default Model

Model used for most requests

Think Model

Model used for complex reasoning tasks

Background Model

Fast model used for simple tasks

WebSearch Model

Model used when web search is needed

Settings

Manage router regex patterns and server settings

Auto-mapping Pattern

Auto-map Regex (e.g., ^claude- to match all Claude models)

Models matching this regex will be routed through WebSearch/Think/Background logic. Leave empty to use default (^claude-).

Background Task Pattern

Background Regex (e.g., (?i)claude.*haiku for Haiku models)

Models matching this regex will use the background model. Leave empty to use default ((?i)claude.*haiku).

OAuth Tokens

Manage your Claude Pro/Max OAuth authentication tokens

Loading OAuth tokens...

No OAuth tokens found

Add a provider with OAuth authentication to see tokens here

Server Actions

Test Models

Test your registered models with custom messages

Select Model

Select Provider (순위 무시)

선택하면 라우팅 순위를 무시하고 해당 provider로 직접 요청합니다

Message

Advanced Options

Max Tokens

Temperature

Enable Streaming

Response

Generating response...

Input Tokens: -

Output Tokens: -

Stop Reason: -

Model: -

Error