# Yibiao Bid Toolbox - AI Bid Proposal Writing Assistant
简体中文 | English
🚀 Out-of-the-box, open-source, and free AI bid proposal writing tool
Yibiao Bid Toolbox is an intelligent bid document creation tool for tendering and bidding scenarios. It is fully open source and includes AI-generated technical proposals, image-and-text generation, commercial bid support, enterprise knowledge base management, duplicate checking, rejection-risk checks, tender information, and more features under development.
It supports all OpenAI-like AI APIs. GPT, DeepSeek, and Volcano Ark are deeply adapted, and local models can also be connected through Ollama, LM Studio, and similar tools.
❓ What problem does it solve?
There are many paid AI bid-writing tools today, but they are extremely expensive. A single proposal can cost tens of yuan, and unless a company reimburses it, small-business workers can hardly afford them. Free tools often have very poor quality. OpenBidKit aims to become the OpenClaw of the bidding field by providing an out-of-the-box, high-quality proposal writing tool. In our tests, generating a 100,000-word bid proposal with DeepSeek V4 Flash costs only about 1 yuan.
AI Proposal Writing · Bid AI · AI Bid Generation · Technical Proposal Drafting · Tender Response Generation
More than proposal draft generation: open-source control, local workspace, reusable knowledge, visual expression, and recoverable workflows.
|
🧩 Open Source & Controllable An open-source AI bid proposal project that can be self-hosted, customized, and adapted to team workflows. |
💻 Local Desktop Workspace Configurations, caches, and generated results are stored locally, suitable for Windows bid-document workflows. |
📄 Multiple Parsing Options Supports local parsing and MinerU parser configuration for both regular documents and more complex files. |
|
📚 Knowledge Base Reuse Store company materials, historical cases, and proposal assets so bid AI output better matches your business context. |
🧩 Images & Diagrams Supports Mermaid preview, generated illustrations, and diagram conversion for Word export. |
🔄 Background Task Recovery Long-running parsing and generation tasks are persisted, so progress can be restored after switching pages. |
|
🛡️ Risk Check Workspaces Duplicate-check and rejection-risk checklist workspaces are reserved for repeated wording and response-completeness checks. |
⚙️ Custom AI Configuration Configure text models, image models, and file parsing providers, with Pi Agent handling all agent tasks. |
✏️ Editable Workflow Outlines, generated content, and expansion results remain editable for human review and final polishing. |
[Watch the usage demo on Bilibili](https://www.bilibili.com/video/BV1sC5i6SE74)
## 🧑💻 Local Development
There is no root-level `package.json` in this repository. The desktop client lives under `client/`, so all client development commands must be run from the `client/` directory.
The client uses Pi Agent exclusively, and the runtime is installed by `npm ci`. Pi's command environment depends on `rg`, `fd`, and `jq`; prepare and verify these tools for your platform before local debugging.
Windows x64:
```powershell
cd client
npm ci
node scripts/prepare-agent-tools.cjs --platform win32 --arch x64
node scripts/verify-agent-tools.cjs --platform win32 --arch x64
npm run dev
```
macOS Apple Silicon:
```bash
cd client
npm ci
node scripts/prepare-agent-tools.cjs --platform darwin --arch arm64
node scripts/verify-agent-tools.cjs --platform darwin --arch arm64
npm run dev
```
macOS Intel:
```bash
cd client
npm ci
node scripts/prepare-agent-tools.cjs --platform darwin --arch x64
node scripts/verify-agent-tools.cjs --platform darwin --arch x64
npm run dev
```
End users who install a GitHub Release package do not need to run these scripts. The release workflow prepares and injects the platform-specific agent command tools automatically. Before building a local installer, run the matching `prepare-agent-tools.cjs` and `verify-agent-tools.cjs` commands.
Regular build verification:
```powershell
cd client
npm run build
```
## 🛠️ Technical Architecture
The current product is an independent desktop client under `client/`. It does not depend on the legacy `frontend/` or `backend/` structure.
- **Desktop**: Electron Main / Preload provides local file access, configuration, export, and background task capabilities.
- **Renderer**: Vite + React + TypeScript, with global CSS and Radix UI primitives.
- **Features**: Technical proposal, knowledge base, duplicate-check workspace, rejection-risk checklist workspace, and settings.
- **Agent Runtime**: Pi Agent uses the text-model configuration, a dedicated AI proxy, bundled command tools, and one global serial queue.
- **Local Data**: Configuration, workspace data, and generated caches are stored under Electron `userData`.
- **Packaging**: Built for Windows / macOS with electron-builder.
### 🏗️ Project Structure
```
Yibiao Bid Toolbox/
├── client/ # Current desktop client
│ ├── electron/ # Main, Preload, IPC, and local services
│ ├── src/ # Renderer source code
│ │ ├── app/ # Routing, menu, and providers
│ │ ├── features/ # Technical proposal, knowledge base, and other modules
│ │ └── shared/ # Shared types, AI helpers, UI, and utilities
│ ├── assets/ # Icons and static assets
│ └── package.json # Client dependencies and packaging config
├── analytics/ # Independent analytics service
├── tools/ # Independent document parsing and MinerU validation tools
└── README.md # Chinese README
```
## 🤝 Contributing
Contributions are welcome.
1. **🐛 Bug Reports**: Report bugs in [Issues](https://github.com/yibiaoai/yibiao-simple/issues).
2. **💡 Feature Requests**: Suggest new features and improvements.
3. **🔧 Code Contributions**: Fork the repository and submit a pull request.
4. **📖 Documentation**: Help improve documentation and usage guides.
## 🍉 Acknowledgements
- Thanks to all users for your support and trust.
- Special thanks to friends from linuxdo for your support and encouragement.
### 🦞 Non-Developer Contributions
People who provide requirement analysis, technical support, test files, useful feedback, free promotion, and other help for this open-source project may not automatically appear in contributors, but they still deserve to be recorded.
|
云峰 |
Engineer X |
专业标书 |
Mr.Erick |
小麦浪的夏天 |
|
韩枫（石化安装培训） |
﹏陌路°天涯 |
刘梦 |
cc |
| - **Official Website**: [https://yibiao.pro](https://yibiao.pro) - **Feedback**: [GitHub Issues](https://github.com/yibiaoai/yibiao-simple/issues) - **Email**: support@yibiao.pro |
|
⭐ If this project helps you, please give it a Star.
⭐ This project has been open sourced and self-recommended in the LINUX DO community. We welcome your supervision, feedback, and contributions.
`AI Proposal Writing` `Bid AI` `AI Bid Generation` `Free Bid Proposal Tool`