Forge

<div align="center"> <h3>One API for all AI models</h3> <p> <a href="https://www.producthunt.com/products/tensorblock-forge?embed=true&utm_source=badge-top-post-badge&utm_medium=badge&utm_source=badge-tensorblock-forge" target="_blank"><img src="https://api.producthunt.com/widgets/embed-image/v1/top-post-badge.svg?post_id=985244&theme=light&period=daily&t=1752002891028" alt="TensorBlock Forge - One API for all AI models | Product Hunt" style="width: 250px; height: 54px;" width="250" height="54" /></a> </p> <p> <a href="#key-features">Features</a> • <a href="#installation">Installation</a> • <a href="#usage">Usage</a> • <a href="#configuration">Configuration</a> • <a href="#security-and-privacy">Security</a> • <a href="#contributing">Contributing</a> • <a href="#license">License</a> </p> </div>

Introduction

Forge is an open-source middleware service that simplifies AI model provider management. It allows you to use multiple AI providers (OpenAI, Anthropic, etc.) through a single, unified API. By storing your provider API keys securely, Forge generates a unified key that works across all your AI applications.

Product Poster

News

We have deployed Forge as an online service, feel free to ✨ learn the service and 🚀 try it now
Forge now enables Claude Code to work with any LLM

Why Forge?

Unified Experience: Use multiple AI models through a single, consistent API
Simplified Key Management: Store provider keys once, use everywhere
Seamless Integration: Compatible with OpenAI API standard for easy integration with existing tools
Enhanced Security: Keys are encrypted and never exposed to end applications

Developer Quickstart

Make your first API request in minutes. Learn the basics of the Forge platform.

cURL Example

curl https://api.forge.tensorblock.co/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $FORGE_API_KEY" \
  -d '{
    "model": "OpenAI/gpt-4o",
    "messages": [
      {
        "role": "developer",
        "content": "You are a helpful assistant."
      },
      {
        "role": "user",
        "content": "Hello!"
      }
    ]
  }'

Python Example

from openai import OpenAI

client = OpenAI(
    base_url="https://api.forge.tensorblock.co/v1", 
    api_key=FORGE_API_KEY,  
)

# models = client.models.list()

completion = client.chat.completions.create(
    model="OpenAI/gpt-4o",
    messages=[
        {"role": "developer", "content": "You are a helpful assistant."},
        {"role": "user", "content": "Hello!"}
    ]
)

print(completion.choices[0].message)

JavaScript/Node.js Example

const { OpenAI } = require("openai");

const client = new OpenAI({
    baseURL: "https://api.forge.tensorblock.co/v1",
    apiKey: FORGE_API_KEY,  
});

// const models = await openai.models.list();

async function main() {
    const completion = await client.chat.completions.create({
        model: "OpenAI/gpt-4o",
        messages: [
            { role: "developer", content: "You are a helpful assistant." },
            { role: "user", content: "Hello!" }
        ]
    });
    console.log(completion.choices[0].message);
}

main();

Key Features

Unified API Key: Store multiple provider API keys and access all with a single Forge API key
OpenAI API Compatible: Drop-in replacement for any application that uses OpenAI's API
Model Mapping: Create custom model names mapped to provider-specific models
Advanced Security: Strong encryption for API keys with JWT-based authentication
High Performance: Built for scalability and concurrent usage
CLI Management: Easy key and user management via included command-line interface
Extensible Architecture: Provider adapter pattern makes it easy to add new AI providers

Installation

Prerequisites

Python 3.12 or newer
UV package manager (installed automatically by the setup scripts)

The project uses UV as the preferred package manager for faster and more reliable dependency installation.

Quick Start

# Clone the repository
git clone https://github.com/yourusername/forge.git
cd forge

# Make sure you have a running PostgreSQL instance.
# You can start one locally using Docker:
# docker compose up -d db
#
# Your DATABASE_URL for the local Docker setup will be:
# DATABASE_URL="postgresql://forge:forge@localhost:5432/forge"

# Create and configure your environment file
cp .env.example .env
# Now, edit the .env file with your specific settings.
# If you used the docker compose command above, you can use the DATABASE_URL provided.
# The `DATABASE_URL` should point to your running PostgreSQL instance.
# Example: DATABASE_URL="postgresql://user:password@localhost/mydatabase"

# Run the setup script
# For Linux/macOS:
chmod +x setup.sh
./setup.sh

# For Windows:
setup.bat

# Start the server
python run.py

Using Docker

# First, create your environment file from the example.
# The Docker container will automatically use this .env file for configuration.
cp .env.example .env
# Now, edit the .env file with your specific settings.

# Build and run with Docker Compose
docker compose up -d

Manual Setup

If you prefer to set up manually:

# After cloning the repo, create your environment file
cp .env.example .env
# Now, edit the .env file with your specific settings.

# Install UV package manager
curl -LsSf https://astral.sh/uv/install.sh | sh  # Linux/macOS
# or
pip install uv  # Windows

# Create a virtual environment
uv venv venv --python=python3.12
source venv/bin/activate  # Linux/macOS
# or
venv\Scripts\activate  # Windows

# Install dependencies
uv pip install -e .
uv pip install --dev -e .

# Run migrations
alembic upgrade head

# Start the server
python run.py

See the detailed installation guide for step-by-step instructions.

Usage

Managing Forge with CLI

Forge comes with a command-line interface for easy management:

# Run in interactive mode
./forge-cli.py

# Or use specific commands
./forge-cli.py register --username myuser --email user@example.com
./forge-cli.py login --username myuser
./forge-cli.py add-key --provider openai --api-key sk-...
./forge-cli.py test --model gpt-4o --message "Hello, AI!"

See all available commands:

./forge-cli.py --help

API Documentation

API documentation is available at /docs when the server is running.

Integration

Connecting Frontends

To use Forge with frontends like CherryStudio, LobeChat, or any OpenAI-compatible application:

Register and add your provider API keys in Forge
Configure your frontend to use Forge's URL and your Forge API key
Use standard model names (e.g., gpt-4o, claude-sonnet-4) or custom mappings

Frontend ➡️ Forge ➡️ AI Provider API

Supported Providers

Forge supports an extensive range of AI providers through both custom adapters and OpenAI-compatible interfaces:

OpenAI
Anthropic
Google Gemini
xAI
DeepSeek
Cohere
Mistral
Nvidia
Alibaba
Fireworks AI
Azure OpenAI
AWS Bedrock
Together AI
OpenRouter
Cerebras
Groq
SambaNova
Moonshot
Hunyuan
Baichuan
Stepfun
01.ai
Nebius
Novita
NScale
DeepInfra
Maritaca
Featherless.ai
Enfer
Inference.net
Kluster.ai
Lambda
Mancer
Redpill.ai
Parasail
Nineteen.ai
Targon
Hyperbolic
SiliconFlow
TensorBlock
Perplexity
Zhipu

The Anthropic integration includes:

Support for Claude 3 models with streaming responses
Automatic model name mapping (e.g., "claude" → "claude-3-opus-20240229")
Message API format support
Token usage reporting

Configuration

Environment Variables

Key environment variables in your .env file, please check .env.example:

# Server settings
HOST=127.0.0.1
PORT=8000

# Database
DATABASE_URL=postgresql://user:password@localhost:5432/forge

# Security
API_KEY_ENCRYPTION_KEY=your_generated_key
JWT_SECRET_KEY=your_generated_key
# Secret key for signing JWTs. Generate a strong, random key.
# Use `openssl rand -hex 32` to generate a key.
SECRET_KEY=
# Secret key for encrypting sensitive data like provider API keys.
# Generate with: python -c "from cryptography.fernet import Fernet; print(Fernet.generate_key().decode())"
ENCRYPTION_KEY=

Model Mapping

Create custom aliases for provider models:

{
  "my-smart-model": "gpt-4o",
  "my-fast-model": "claude-instant"
}

Security and Privacy

At Forge, we take the security and privacy of your API keys very seriously. Since our service requires storing provider API keys (like OpenAI, Anthropic, etc.), we've implemented multiple layers of protection:

Forge

Install / Use

README

Forge

Introduction

Developer Quickstart

cURL Example

Python Example

JavaScript/Node.js Example

Key Features

Installation

Prerequisites

Quick Start

Using Docker

Manual Setup

Usage

Managing Forge with CLI

API Documentation

Integration

Connecting Frontends

Supported Providers

Configuration

Environment Variables

Model Mapping

Security and Privacy

API Key Encryptio