Content
[](README.md)
[](README_CN.md)
Nexent is an open-source agent SDK and platform that transforms natural language describing processes into complete multimodal services—without orchestration and complex drag-and-drop. Built on the MCP (Model Control Protocol) tool ecosystem, Nexent offers flexible model integration, scalable data processing, and powerful knowledge base management. Our goal is simple: to integrate data, models, and tools into an intelligent hub, enabling anyone to easily incorporate Nexent into their projects, making daily workflows smarter and more interconnected.
> One prompt, infinite possibilities.
### 🌟 Experience Nexent Now
- 🌐 Visit our [official website](http://nexent.tech/) for more information
- 🚀 [Try it out](http://nexent.tech/try) to experience the powerful features of Nexent
https://github.com/user-attachments/assets/0758629c-3477-4cd4-a737-0aab330d53a7
# 🤝 Join Our Community
> *If you want to go fast, go alone; if you want to go far, go together.*
We are still in the initial open-source phase, aiming to release the stable version of Nexent v1 by **June 2025**. Until then, we will continue to rapidly release core features—and we look forward to your participation:
* **🗺️ Check out our [feature roadmap](https://github.com/orgs/ModelEngine-Group/projects/6)** to explore current and upcoming features.
* **🔍 Try the current version** and leave your thoughts or report bugs in the [issue tracker](https://github.com/ModelEngine-Group/nexent/issues).
> *Rome wasn't built in a day.*
While many key features are still under active development, if our vision resonates with you, please join us through the **[contribution guidelines](CONTRIBUTING_CN.md)** to help shape Nexent.
Early contributors will not be overlooked: from special badges and memorabilia to other substantial rewards, we are committed to thanking those pioneers who helped bring Nexent to life.
Most importantly, we need visibility. Please star ⭐ the repository and follow us, share with friends, and help more developers discover Nexent—every click brings new participants to the project and keeps the momentum going.
# ⚡ Get Started
### 1. 📋 System Requirements
| Resource | Minimum Requirements |
|----------|---------------------|
| **CPU** | 2 cores |
| **Memory** | 6 GiB |
| **Software** | Docker and Docker Compose installed |
### 2. 🛠️ Quick Start with Docker Compose
```bash
git clone https://github.com/ModelEngine-Group/nexent.git
cd nexent/docker
cp .env.example .env # fill only necessary configs
bash deploy.sh
```
Once the container is running, open **http://localhost:3000** in your browser and follow the setup wizard.
### 3. 🤖 Model Configuration and Recommended Providers
We recommend using the following model providers:
| Model Type | Provider | Description |
|------------|----------|-------------|
| LLM and VLLM | [Silicon Flow](https://siliconflow.cn/) | Offers free credits |
| LLM and VLLM | [Alibaba Cloud Bailian](https://bailian.console.aliyun.com/) | Offers free credits |
| Embedding | [Jina](https://jina.ai/) | Offers free credits |
| TTS and STT | [Volcano Engine Voice](https://www.volcengine.com/product/voice-tech) | Free for individual users |
| Search | [EXA](https://exa.ai/) | Offers free credits |
You will need to enter the following information on the model configuration page:
- Base URL
- API Key
- Model Name
The following configurations need to be added to your `.env` file (we will front-end these configurations as soon as possible):
- TTS and STT related configurations
- EXA Search API Key
> ℹ️ Due to tight development, we currently only support Jina Embedding models. Support for other models will be added in future versions. For obtaining the Jina API key, please refer to our [FAQ](FAQ_CN.md).
### 4. ❓ Need Help?
- Browse the [FAQ](FAQ_CN.md) for common installation issues.
- Ask questions in our [Discord community](https://discord.gg/tb5H3S3wyv).
- Submit bug reports or feature suggestions in [GitHub Issues](https://github.com/ModelEngine-Group/nexent/issues).
### 5. 🔧 Developing Nexent
Want to build from source or add new features? Check out the [contribution guidelines](CONTRIBUTING_CN.md) for step-by-step instructions.
### 6. 🛠️ Building from Source
Want to run Nexent from source? Check our [developer guide](DEVELOPPER_NOTE_CN.md) for detailed setup instructions and customization options.
## ✨ Key Features
`1` **Automatic Prompt Generation for Agents**
Transforms natural language into prompts executable by the Agent. Nexent can automatically select the right tools based on your needs and plan the best execution path for each request.
`2` **Scalable Data Processing Engine**
Supports fast OCR and table structure extraction for 20+ data formats, smoothly scaling from single-process to large-scale batch processing pipelines.
`3` **Personal Knowledge Base**
Real-time file import and automatic summarization allow agents to access personal and global knowledge instantly, understanding what each knowledge base can provide.
`4` **Internet Knowledge Search**
Connects with 5+ web search providers, enabling agents to combine the latest internet information with your private data.
`5` **Knowledge-Level Traceability**
Provides precise citations from web and knowledge base sources, making every fact verifiable.
`6` **Multimodal Understanding and Dialogue**
Speak, type, upload files, or display images. Nexent understands voice, text, and images, and can even generate new images as needed.
`7` **MCP Tool Ecosystem**
Insert or build Python plugins that comply with MCP specifications; switch models, tools, and chains without modifying core code.
# 🐛 Known Issues
1. 🔄 **Knowledge Base Refresh Delay**
We are aware that the knowledge base refresh mechanism currently has some delays, occasional errors, or refresh failures. We plan to refactor this part soon, but please note that this is merely a logical issue in task management, and the actual data processing speed is unaffected.
2. 🤖 **Limited Model Provider Support**
Our current support for different model providers, including voice and multimodal models, is limited. We will rapidly update this support in the coming weeks, so stay tuned!
3. 🎨 **Lack of Agent Creation and Configuration Interface**
We have noted that there is currently no visual interface for agent creation, tool selection, and configuration. We plan to implement this feature in the coming weeks to make the agent creation process more intuitive and user-friendly.
4. 📦 **Large Docker Image Size**
We have noticed that the current Docker image size is large due to the inclusion of middleware components from the development process. We are actively optimizing and reducing the image size to enable faster and lighter deployments.
# 👥 Contribution Guidelines
We welcome all forms of contributions! Whether it's fixing bugs, adding features, or improving documentation, your help can make Nexent better.
- 📖 Read our [contribution guidelines](CONTRIBUTING_CN.md) to start contributing
- 🐛 Report bugs or suggest features in [GitHub Issues](https://github.com/ModelEngine-Group/nexent/issues)
- 💬 Join our [Discord community](https://discord.gg/tb5H3S3wyv) to discuss ideas
# 💬 Community and Contact
Join our [Discord community](https://discord.gg/tb5H3S3wyv) to connect with other developers and get help!
# 📄 License
Nexent is licensed under the [Apache License 2.0](LICENSE) with additional conditions. Please read the [LICENSE](LICENSE) file for details.
