中文 | English

_This project is a secondary development based on [one-hub](https://github.com/MartialBE/one-api)_

^{AI Wave Community} · ^{(Offering public API and AI bots in-group)} ### [📚 View Original Project Documentation](https://one-hub-doc.vercel.app/)

## Current Differences from the Original Version (Latest Image) - Supports **batch channel deletion** - Supports `LinuxDo` login - Supports **dark mode following system settings** - Supports **batch adding models to multiple channels** - Supports configuring **case-insensitive model names** - Supports configuring **unified request-response model names** - Supports **removing specified parameters** from channel extra parameters - Supports **model variable substitution** in channel `BaseURL` - Supports **pass-through of extra parameters** for native `/gemini` image generation requests - Supports **pass-through of extra parameters** for `Claude` channels (both OpenAI format and native format) - Supports **custom channels** using native `Claude` routing - integrating `ClaudeCode` - Supports `VertexAI` channels using native `Gemini` routing - integrating `GeminiCli` - Supports `VertexAI` channels using native `Claude` routing - integrating `ClaudeCode` - Supports configuring multiple `Regions` for `VertexAI` channels, with random `Region` selection per request - Supports native **video generation requests** (`Veo` series models) for `Google Gemini` channels via `/gemini` - Supports `gemini-2.0-flash-preview-image-generation` for text-to-image/image-to-image, compatible with `OpenAI` chat interfaces - Added **invitation code settings** module - Added **user group functionality for batch channel addition** - Added **configuration for whether empty responses are billed** (default: billed) - Added **time period conditions in analytics - top-up statistics** - Added **RPM / TPM / CPM display in analytics** - Added **invitation top-up rebate feature** (optional types: fixed/percentage) - Fixed `bug` where user-related APIs were ineffective - Fixed `bug` with missing fields in invitation records - Fixed `bug` where hardcoded timezones affected statistical data - Fixed `bug` with payment callback exceptions in multi-instance deployments - Fixed `bug` with floating-point `token` calculations for Zhipu `GLM` models - Fixed `bug` where API routing allowed `CDN` caching, leading to privilege escalation - Fixed several `bugs` where user quota cache inconsistencies with `DB` data caused billing anomalies - Removed meaningless original price-related styles from log functionality - Optimized email rule validation - Optimized several `UI` interactions - Optimized logic for disabled channel email notifications - Optimized `VertexAI` authentication caching - ... ## Deployment > Follow the original deployment tutorial and replace the image with `deanxv/done-hub`. > Database-compatible; the original version can directly pull this image for migration. ## Acknowledgments - This program uses the following open-source projects: - [one-hub](https://github.com/MartialBE/one-api) as the foundation of this project Thanks to the authors and contributors of the above projects.