🐾 NyaProxy - Universal API Proxy

The smart way to balance, secure, and monitor all your API interactions

🌈 Introduction

Warning

This project is currently under active development. Documentation may not reflect the latest changes. If you encounter unexpected behavior, please consider using a previous stable version or report issues on our GitHub repository.

NyaProxy acts like a smart, central manager for accessing various online services (APIs) – think AI tools (like OpenAI, Gemini, Anthropic), image generators, or almost any web service that uses access keys. It helps you use these services more reliably, efficiently, and securely.

Here's how NyaProxy can help:

• Share the Load: Automatically spreads requests across multiple access keys so no single key gets overwhelmed (Load Balancing).
• Stay Online: If one key fails, NyaProxy automatically tries another, keeping your application running smoothly (Failover/Resilience).
• Save Costs: Optimizes how your keys are used, potentially lowering your bills.
• Boost Security: Hides your actual access keys behind the proxy, adding a layer of protection.
• Track Usage: Provides a clear dashboard to see how your keys and services are being used in real-time.

🌟 Core Capabilities

Feature	Description	Config Reference
🔄 Token Rotation	Automatic key cycling across multiple providers	variables.keys
⚖️ Load Balancing	5 strategies: Round Robin, Random, Least Request, Fastest Response, Weighted	load_balancing_strategy
🚦 Rate Limiting	Granular controls per endpoint/key with smart queuing	rate_limit
🕵️ Request Masking	Dynamic header substitution across multiple identity providers	headers + variables
📊 Real-time Metrics	Interactive dashboard with request analytics and system health	dashboard
🔧 Body Substitution	Dynamic JSON payload transformation using JSONPath	request_body_substitution

📥 Quick Start

One-Click Deployment (No Fuss, No Muss!)

Pick your favorite platform and let's go!

Deploy to Render

Deploy to Railway

Note

NyaProxy automatically creates a basic working configuration when it starts. You just need to access the /config endpoint to add your API keys!

Tip

You can use Gemini AI Studio to get a free API key for testing. Gemini's API is OpenAI-compatible and works seamlessly with NyaProxy. Get a Gemini API key here.

Local Deployment (For the DIY Enthusiasts!)

Prerequisites

• Python 3.10 or higher
• Docker (optional, for containerized deployment)

Installation

1. Install from PyPI (the easiest way!)

pip install nya-proxy

2. Run NyaProxy

nyaproxy

...or provide your own config file:

nyaproxy --config config.yaml

3. Verify Your Setup

Visit http://localhost:8080/config to access the configuration UI.

Important

If you expose this proxy to the internet, make sure to set a strong API key in your configuration to prevent unauthorized access. The first key in your API keys array will be used as the master key for accessing sensitive interfaces like the dashboard and configuration UI, while additional keys can be used for regular API requests only.

If no master API key is specified, no login page will be shown and anyone can access the dashboard and configuration UI. This is convenient for local testing but not recommended for production environments.

Check out http://localhost:8080/dashboard for the snazzy management dashboard with all your API traffic visualizations.

Detailed Deployment Guides

For step-by-step instructions tailored to beginners, check out our detailed deployment guides:

• Docker Deployment Guide - Run with Docker or Docker Compose
• PIP Installation Guide - Direct Python installation

Install from Source

# Clone the repository
git clone https://github.com/Nya-Foundation/nyaproxy.git
cd nyaproxy

# Install dependencies
pip install -e .

# Run NyaProxy
nyaproxy

Docker

docker run -d \
  -p 8080:8080 \
  # -v ${PWD}/config.yaml:/app/config.yaml \
  # -v ${PWD}/app.log:/app/app.log \
  k3scat/nya-proxy:latest

Configuration

Configuration reference can be found under Configs folder folder

# Basic config.yaml example for OpenAI API Compatible Endpoints
# NyaProxy Configuration File
# This file contains server settings and API endpoint configurations

server:
  api_key: 
  logging:
    enabled: true
    level: debug
    log_file: app.log
  proxy:
    enabled: false
    address: socks5://username:password@proxy.example.com:1080
  dashboard:
    enabled: true
  cors:
    # Allow all origins with "*", but specify exact origins when allow_credentials is true for security
    allow_origins: ["*"]
    allow_credentials: true
    allow_methods: ["GET", "POST", "PUT", "DELETE", "OPTIONS"]
    allow_headers: ["*"]

# Default configuration applied to all API endpoints unless overridden
default_settings:
  key_variable: keys
  key_concurrency: true # Enable concurrent requests per API key for better performance
  randomness: 0.0 # Random delay of (0.0-x)s to introduce variability in request timing and avoid detection
  load_balancing_strategy: round_robin
  queue:
    max_size: 200
    expiry_seconds: 300
  rate_limit:
    enabled: true
    endpoint_rate_limit: 10/s # Default endpoint rate limit - can be overridden per API
    key_rate_limit: 10/m # Default key rate limit - can be overridden per API
    ip_rate_limit: 1000/d # IP-based rate limit to protect against abuse and key redistribution
    user_rate_limit: 1000/d # User-based rate limit per proxy API key defined in server section
    rate_limit_paths: 
      - "*"
  retry:
    enabled: true
    mode: key_rotation
    attempts: 3
    retry_after_seconds: 1
    retry_request_methods: [ POST, GET, PUT, DELETE, PATCH, OPTIONS ]
    retry_status_codes: [ 429, 500, 502, 503, 504 ]
  timeouts:
    request_timeout_seconds: 300

  # Request body substitution settings
  request_body_substitution:
    enabled: false
    # Substitution rules for request body with JMEPath
    rules:
      - name: "Remove frequency_penalty"
        operation: remove
        path: "frequency_penalty"
        conditions:
          - field: "frequency_penalty"
            operator: "exists"
      - name: "Remove presence_penalty"
        operation: remove
        path: "presence_penalty"
        conditions:
          - field: "presence_penalty"
            operator: "exists"
apis:
  gemini:
    # OpenAI-compatible API endpoint
    name: Google Gemini API
    # Supported endpoints:
    # Gemini: https://generativelanguage.googleapis.com/v1beta/openai
    # OpenAI: https://api.openai.com/v1
    # Anthropic: https://api.anthropic.com/v1
    # DeepSeek: https://api.deepseek.com/v1
    # Mistral: https://api.mistral.ai/v1
    # OpenRouter: https://api.openrouter.ai/v1
    # Ollama: http://localhost:11434/v1
    endpoint: https://generativelanguage.googleapis.com/v1beta/openai
    aliases:
    - /gemini
    key_variable: keys
    headers:
      Authorization: 'Bearer ${{keys}}'
    variables:
      keys:
      - your_gemini_key_1
      - your_gemini_key_2
      - your_gemini_key_3
    load_balancing_strategy: least_requests
    rate_limit:
      enabled: true
      # For Gemini-2.5-pro-exp-03-25, rate limits per key are 5 RPM and 25 RPD
      # Endpoint rate limit should be n × Per-Key-RPD, where n is the number of keys
      endpoint_rate_limit: 75/d
      key_rate_limit: 5/m
      # Paths to apply rate limits (regex supported) - defaults to all paths "*"
      rate_limit_paths:
        - "/chat/*"
        - "/images/*"

  test:
    name: Test API
    endpoint: http://127.0.0.1:8082
    key_variable: keys
    randomness: 5
    headers:
      Authorization: 'Bearer ${{keys}}'
      User-Agent: ${{agents}}
    variables:
      keys:
      - your_test_key_1
      - your_test_key_2
      - your_test_key_3
      agents:
      - test_agent_1
      - test_agent_2
      - test_agent_3
    load_balancing_strategy: least_requests
    rate_limit:
      enabled: true
      endpoint_rate_limit: 20/m
      key_rate_limit: 5/m
      ip_rate_limit: 1000/d
      user_rate_limit: 1000/d
      rate_limit_paths:
        - "/v1/*"

  # feel free to add more APIs here, just follow the same structure as above

📡 Service Endpoints

Service	Endpoint	Description
API Proxy	http://localhost:8080/api/<endpoint_name>	Main proxy endpoint for API requests
Dashboard	http://localhost:8080/dashboard	Real-time metrics and monitoring
Config UI	http://localhost:8080/config	Visual configuration interface

Note

Replace 8080 and localhost with your configured port and host setting if different

🔧 API Configuration

OpenAI-Compatible APIs (Gemini, Anthropic, etc)

gemini:
  name: Google Gemini API
  endpoint: https://generativelanguage.googleapis.com/v1beta/openai
  aliases:
    - /gemini
  key_variable: keys
  headers:
    Authorization: 'Bearer ${{keys}}'
  variables:
    keys:
      - your_gemini_key_1
      - your_gemini_key_2
  load_balancing_strategy: least_requests
  rate_limit:
    endpoint_rate_limit: 75/d     # Total endpoint limit
    key_rate_limit: 5/m          # Per-key limit
    rate_limit_paths:
      - "/chat/*"            # Apply limits to specific paths
      - "/images/*"

Generic REST APIs

novelai:
  name: NovelAI API
  endpoint: https://image.novelai.net
  aliases:
    - /novelai
  key_variable: tokens
  headers:
    Authorization: 'Bearer ${{tokens}}'
  variables:
    tokens:
      - your_novelai_token_1
      - your_novelai_token_2
  load_balancing_strategy: round_robin
  rate_limit:
    endpoint_rate_limit: 10/s
    key_rate_limit: 2/s

🔒 Security Features

Multiple API Keys Support

NyaProxy supports using multiple API keys for authentication:

server:
  api_key: 
    - your_master_key_for_admin_access
    - another_api_key_for_proxy_only
    - yet_another_api_key_for_proxy_only

Tip

The first key in the list acts as the master key with full access to the dashboard and configuration UI. Additional keys can only be used for API proxy requests. This enables you to share limited access with different teams or services.

Caution

When sharing your NyaProxy instance, never share your master key. Instead, create additional keys for different users or applications.

Advanced Features

Dynamic Header Substitution

NyaProxy's powerful templating system allows you to create dynamic headers with variable substitution:

apis:
  my_api:
    headers:
      Authorization: 'Bearer ${{keys}}'
      X-Custom-Header: '${{custom_variables}}'
    variables:
      keys:
        - key1
        - key2
      custom_variables:
        - value1
        - value2

Note

Variables in headers are automatically substituted with values from your variables list, following your configured load balancing strategy.

Use cases include:

• Rotating between different authentication tokens
• Cycling through user agents to avoid detection
• Alternating between different account identifiers

Request Body Substitution

Dynamically transform JSON payloads using JMESPath expressions to add, replace, or remove fields:

request_body_substitution:
  enabled: true
  rules:
    - name: "Default to GPT-4"
      operation: set
      path: "model"
      value: "gpt-4"
      conditions:
        - field: "model"
          operator: "exists"

For detailed configuration options and examples, see the Request Body Substitution Guide.

🖥️ Management Interfaces

Real-time Metrics Dashboard

Monitor at http://localhost:8080/dashboard:

• Request volumes and response times
• Rate limit status and queue depth
• Key usage and performance metrics
• Error rates and status codes

Visual Configuration Interface

Manage at http://localhost:8080/config:

• Live configuration editing
• Syntax validation
• Variable management
• Rate limit adjustments
• Auto reload on save

🛡️ Reference Architecture

graph TD
    A[Client] --> B[Nginx]
    B --> C[NyaProxyAuth]
    C --> D[NyaProxyApp]
    D --> E[API Providers]
    F[Monitoring] --> D

Loading

🌌 Future Roadmap

graph LR
A[Q1 2025] --> B[📡 Documentation Enhancement ]
A --> C[🔄 Border Test Coverage ]
B --> D[📈 API Key Usage/Balance Tracking]
C --> E[📊 UI/UX Enhancement ]
F[Q2 2025] --> G[🧩 Plugin System]
F --> H[🔍 Custom Metrics API]

Loading

❤️ Community

Note

Need support? Contact k3scat@gmail.com or join our discord community at Nya Foundation

📈 Project Growth