akasha-whisper is a Golang project offering a unified, user-friendly API for integrating and interacting with multiple AI models.
While similar to one-api, akasha-whisper provides additional features:
- Dynamic client weight configuration and intelligent load balancing.
- Support for multiple providers for each model, allowing integration of various sources like OpenAI and Azure, with automatic provider selection based on predefined weight and pricing configurations.
For detailed documentation, visit the Akasha-whisper documentation.
| Name | Description | URL | Status |
|---|---|---|---|
| Chat | complete chat with given prompt | v1/chat/completions |
✅ |
| Model | list available models | /v1/models |
✅ |
| Embeddings | get embeddings for given text | /v1/embeddings |
✅ |
| Image | generate image with given prompt | v1/images/generations |
WIP |
| Speech | generate speech audio from given text | v1/audio/speech |
WIP |
| Transcription | generate text from given audio file | v1/audio/transcription |
WIP |
The following models and providers have been thoroughly tested by the development team and confirmed to function reliably.
| Name | Provider | Home Page | BaseURL |
|---|---|---|---|
| GPT | OpenAI | OpenAI | https://api.openai.com/v1 |
| GLM | ZhipuAI | ZhipuAI | https://open.bigmodel.cn/api/paas/v4 |
| qwen | Alibaba | Tongyi | https://dashscope.aliyuncs.com/compatible-mode/v1 |
| hunyuan | Tencent | Hunyuan | https://api.hunyuan.cloud.tencent.com/v1 |
Thanks to JetBrains for providing Open Source development license(s) for this project.