TokenFactory

AI gateway you self-host: route many model providers through one API surface, with users, keys, quotas, and an admin UI.

---

Upstream (required reading)

TokenFactory is derived from the QuantumNous/new-api project (“New API”). That repository is the authoritative upstream for design, protocol coverage, and community history. This fork may diverge; for behavior and APIs, treat upstream docs as the baseline and verify against your build.

Upstream repository	github.com/QuantumNous/new-api
License	GNU AGPL v3.0 — same for modifications here; see NOTICE
Network use	If you offer a modified version over a network to others, AGPL-3.0 section 13 requires you to provide the corresponding full source under the same license.

---

简体中文 | 繁體中文 | English | Français | 日本語

   

Quick Start • Documentation • Languages • Deployment • License • Help

About this repository

TokenFactory is a self-hosted control plane for aggregating upstream AI vendors: one place to configure channels, map models, enforce access, and observe usage. You run the binary (or container), point clients at it, and manage everything from the web console.

This README describes this fork’s packaging and pointers. It does not replace the upstream feature list or legal notices—those remain tied to QuantumNous/new-api and the license files in this tree.

Compliance & disclaimer

• Use only in line with provider terms (e.g. OpenAI Terms of Use) and applicable law. No illegal or abusive use.
• In China, follow registration and compliance rules for generative AI services (e.g. 《生成式人工智能服务管理暂行办法》); do not offer unregistered public generative AI services where prohibited.
• No warranty: treat this as self-supported infrastructure unless you arrange your own support.

---

Quick Start

Using Docker Compose (Recommended)

# Clone the project
git clone https://github.com/QuantumNous/token-factory.git
cd token-factory

# Edit docker-compose.yml configuration
nano docker-compose.yml

# Start the service
docker-compose up -d

Using Docker Commands

# Pull the latest image
docker pull ghcr.io/fyinfor/token-factory:latest

# Using SQLite (default)
docker run --name token-factory -d --restart always \
  -p 3000:3000 \
  -e TZ=Asia/Shanghai \
  -v ./data:/data \
  ghcr.io/fyinfor/token-factory:latest

# Using MySQL
docker run --name token-factory -d --restart always \
  -p 3000:3000 \
  -e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi" \
  -e TZ=Asia/Shanghai \
  -v ./data:/data \
  ghcr.io/fyinfor/token-factory:latest

💡 Tip: -v ./data:/data will save data in the data folder of the current directory, you can also change it to an absolute path like -v /your/custom/path:/data

---

When the stack is healthy, open http://localhost:3000. More install paths (bare metal, panels, etc.): installation docs.

---

Documentation

The QuantumNous/new-api ecosystem publishes the reference manuals for APIs, models, and operations. TokenFactory tracks that stack; use the docs as the source of truth and validate against your build.

Manual (EN / ZH)	docs.newapi.pro — English · 简体中文
Environment variables	Configuration reference
Relay / REST API	API documentation
Feature overview	Features introduction
FAQ & community	FAQ · Channels
Deep dive (third-party)	DeepWiki — QuantumNous/new-api

This repository: report fork-specific bugs (packaging, defaults, CI) here. If the behavior matches upstream, reproduce on QuantumNous/new-api and follow their contribution guidelines.

---

What you get (at a glance)

Capabilities come from the upstream codebase; this is a summary, not an exhaustive spec:

• Relay — many vendor adapters behind a unified API surface (OpenAI-compatible and other formats per upstream).
• Console — channels, model mapping, users, keys, usage and billing configuration.
• Policies — quotas, rate limits, retries, and optional cache when Redis is enabled.
• Storage — SQLite, MySQL, or PostgreSQL; optional Redis for sessions/cache/crypto as documented upstream.

For model-by-model and endpoint-by-endpoint detail, use the API docs and QuantumNous/new-api releases.

---

Supported languages

Code	Language
zh-CN	Chinese (Simplified)
zh-TW	Chinese (Traditional)
en	English
fr	French
ru	Russian
ja	Japanese
vi	Vietnamese
id	Indonesian
ms	Malay
th	Thai
sw	Swahili

---

Deployment

[!TIP] Latest Docker image: ghcr.io/fyinfor/token-factory:latest

📋 Deployment Requirements

Component	Requirement
Local database	SQLite (Docker must mount /data directory)
Remote database	MySQL ≥ 5.7.8 or PostgreSQL ≥ 9.6
Container engine	Docker / Docker Compose

⚙️ Environment Variable Configuration

Common environment variable configuration

Variable Name	Description	Default Value
SESSION_SECRET	Session secret (required for multi-machine deployment)	-
CRYPTO_SECRET	Encryption secret (required for Redis)	-
SQL_DSN	Database connection string	-
REDIS_CONN_STRING	Redis connection string	-
STREAMING_TIMEOUT	Streaming timeout (seconds)	300
STREAM_SCANNER_MAX_BUFFER_MB	Max per-line buffer (MB) for the stream scanner; increase when upstream sends huge image/base64 payloads	64
MAX_REQUEST_BODY_MB	Max request body size (MB, counted after decompression; prevents huge requests/zip bombs from exhausting memory). Exceeding it returns 413	32
AZURE_DEFAULT_API_VERSION	Azure API version	2025-04-01-preview
ERROR_LOG_ENABLED	Error log switch	false
PYROSCOPE_URL	Pyroscope server address	-
PYROSCOPE_APP_NAME	Pyroscope application name	token-factory
PYROSCOPE_BASIC_AUTH_USER	Pyroscope basic auth user	-
PYROSCOPE_BASIC_AUTH_PASSWORD	Pyroscope basic auth password	-
PYROSCOPE_MUTEX_RATE	Pyroscope mutex sampling rate	5
PYROSCOPE_BLOCK_RATE	Pyroscope block sampling rate	5
HOSTNAME	Hostname tag for Pyroscope	token-factory

📖 Complete configuration: Environment Variables Documentation

🔧 Deployment Methods

Docker Compose: use the Quick Start commands above (clone → edit docker-compose.yml → docker-compose up -d).

Alternative: plain Docker run

Using SQLite:

docker run --name token-factory -d --restart always \
  -p 3000:3000 \
  -e TZ=Asia/Shanghai \
  -v ./data:/data \
  ghcr.io/fyinfor/token-factory:latest

Using MySQL:

docker run --name token-factory -d --restart always \
  -p 3000:3000 \
  -e SQL_DSN="root:123456@tcp(localhost:3306)/oneapi" \
  -e TZ=Asia/Shanghai \
  -v ./data:/data \
  ghcr.io/fyinfor/token-factory:latest

💡 Path explanation:

• ./data:/data - Relative path, data saved in the data folder of the current directory
• You can also use absolute path, e.g.: /your/custom/path:/data

BaoTa Panel

1. Install BaoTa Panel (≥ 9.2.0 version)
2. Search for TokenFactory in the application store
3. One-click installation

📖 Tutorial with images

⚠️ Multi-machine Deployment Considerations

[!WARNING]

• Must set SESSION_SECRET - Otherwise login status inconsistent
• Shared Redis must set CRYPTO_SECRET - Otherwise data cannot be decrypted

🔄 Channel Retry and Cache

Retry configuration: Settings → Operation Settings → General Settings → Failure Retry Count

Cache configuration:

• REDIS_CONN_STRING: Redis cache (recommended)
• MEMORY_CACHE_ENABLED: Memory cache

---

Lineage

Repository	Role
QuantumNous/new-api	Upstream — New API (AGPL-3.0). Start here for history, issues that are not fork-specific, and feature design.
One API	Earlier MIT-licensed codebase in the same family tree.
Midjourney-Proxy	Optional Midjourney integration (see upstream docs).

Tools maintained around the ecosystem (e.g. neko-api-key-tool) are documented upstream.

---

Help

📖 Documentation Resources

Resource	Link
📘 FAQ	FAQ
💬 Community Interaction	Communication Channels
🐛 Issue Feedback	Issue Feedback
📚 Complete Documentation	Official Documentation

🤝 Contribution Guide

Welcome all forms of contribution!

• 🐛 Report Bugs
• 💡 Propose New Features
• 📝 Improve Documentation
• 🔧 Submit Code

---

License

This project (TokenFactory) is licensed under the GNU Affero General Public License v3.0 (AGPLv3). Modifications and further derivatives remain under AGPL-3.0 unless you obtain a separate commercial license from the copyright holders.

Attribution: TokenFactory is derived from QuantumNous/new-api (New API), which is also under AGPL-3.0. The project chain includes One API (MIT License) as an earlier base. Please retain upstream notices and this repository’s LICENSE and NOTICE. Under AGPL-3.0 section 13, if you run a modified version as a network service for others, you must offer them the corresponding complete source code under the same license.

If your organization's policies do not permit the use of AGPLv3-licensed software, or if you wish to avoid the open-source obligations of AGPLv3, please contact us at: support@quantumnous.com

---

TokenFactory — self-hosted AI gateway (this fork).

Upstream: QuantumNous/new-api · Docs: docs.newapi.pro · This repo: issues

_{The New API project is developed by QuantumNous and contributors. JetBrains supports open-source development through free IDE licenses.}