English | Simplified Chinese | Japanese
FastGPT is a knowledge base Q&A system based on the LLM large language model, providing out-of-the-box data processing, model calling, and other capabilities. It can also orchestrate workflows visually through Flow to implement complex Q&A scenarios!
fastgpt.mp4
- ๐ Temporarily available in China: fastgpt.in
- ๐ International version: fastgpt.run
![]() |
![]() |
![]() |
![]() |
1
Application Orchestration Capabilities
- Provide simple mode, no need for manual orchestration
- Guided conversation steps
- Workflow orchestration
- Source file reference tracking
- Module encapsulation, enabling multi-level reuse
- Hybrid retrieval & re-ranking
- Tool module
- Embed Laf to create HTTP modules online
- Plugin encapsulation functionality
2
Knowledge Base Capabilities
- Multi-library reuse, mixed usage
- Chunk recording modification and deletion
- Support for setting vector models for knowledge bases individually
- Source file storage
- Support manual input, direct segmentation, QA split import
- Support pdf, docx, txt, html, md, csv
- Support URL reading, CSV batch import
- Support PPT, Excel import
- Support file reader
- More data preprocessing solutions
3
Application Debugging Capabilities
- Single-point search test in the knowledge base
- Feedback references during conversation and can modify or delete
- Full context presentation
- Complete presentation of intermediate values in modules
- Advanced orchestration DeBug mode
4
OpenAPI Interface
- completions interface (chat mode aligned with GPT interface)
- CRUD for knowledge base
- CRUD for conversation
5
Operational Capabilities
- Share window without login
- Iframe one-click embedding
- Chat window embedding supports customizing Icon, default opening, dragging, etc.
- Unified consultation of conversation records and data annotation
Project technology stack: NextJs + TS + ChakraUI + Mongo + Postgres (Vector plugin)
-
โก Quick Deployment
The server of Sealos is abroad, no extra network issues to handle, no server, no magic, no domain name required, supports high concurrency & dynamic scaling. Click the button below for one-click deployment ๐
As it involves deploying a database, it will take 2~4 minutes to access normally after deployment. The lowest configuration is used by default, so it might be slow on first access. For related tutorials, see: Deploying FastGPT with Sealos
- Quick Start Local Development
- Deploying FastGPT
- System Configuration File Explanation
- Multi-model Configuration
- Version Update/Upgrade Introduction
- OpenAPI API Documentation
- Detailed Explanation of Knowledge Base Structure
Join by adding wx assistant:
- Laf: Quickly integrate third-party applications in 3 minutes
- Sealos: Quickly deploy cluster applications
- One API: Multi-model management, supports Azure, Wenxin Yiyan, etc.
- TuShan: Set up a backend management system in 5 minutes
We warmly welcome contributions of all forms. If you're interested in contributing code, you can check our GitHub Issues to showcase your ideas and talents.
|
|
---|---|
|
|
|
This repository follows the FastGPT Open Source License open source license.
- Allowed for direct commercial use as a backend service, but not allowed to provide SaaS services.
- Any form of commercial service without commercial authorization must retain related copyright information.
- For full details, please see FastGPT Open Source License
- Contact: [email protected], Click to view commercial pricing strategy