The history and evolution of FastGPT, from its inception to becoming a leading knowledge-based AI Agent platform.
FastGPT is a knowledge-based AI Agent platform built on LLMs with RAG capabilities, visual workflow orchestration, and bidirectional MCP support. Developed by Sealos (LanHuan/Labring), it has become one of the most popular platforms for enterprise AI agents with over 27,000 GitHub stars.
FastGPT was created by Sealos (also known as LanHuan or Labring), a cloud-native infrastructure company. The project emerged to address the need for:
- Easy-to-deploy RAG systems
- Visual workflow orchestration for AI
- Enterprise-ready AI agent platform
- Multi-format knowledge base support
The founding mission was to provide a AI Agent platform that:
- Simplifies RAG application development
- Provides visual workflow orchestration
- Supports enterprise deployment
- Maintains open-source values
- Initial Release - First public version of FastGPT
- Core Features - Basic RAG, knowledge base, simple workflows
- Platform Support - Docker deployment
- Database - MongoDB integration
¶ 2024 - Growth and Expansion
Early 2024:
- Visual Workflow - Drag-and-drop workflow builder
- OpenAPI - GPT-compatible API endpoints
- Voice Support - Voice input/output capabilities
- User Base - Thousands of deployments
Mid 2024:
- Hybrid Search - Vector + full-text search
- Reranking - BGE-Reranker integration
- Multi-Format Support - PDF, DOCX, CSV, URL
- Sealos Cloud - One-click deployment
Late 2024:
- MCP Support - Model Context Protocol integration
- Workflow Plugins - Reusable plugin system
- Operations Dashboard - Conversation review and analytics
Early 2025:
- Agent Mode - Planning agent for complex tasks
- Team Collaboration - Multi-user support
- API Key Management - Enhanced API security
- License Update - Apache-2.0 with additional conditions
Mid 2025:
- LLM Request Tracking - Retain all LLM requests/responses
- Chat Logs - Enhanced logging and filtering
- Template Marketplace - Pre-built workflow templates
- Context-Based Agent - Long task decomposition
Late 2025:
- Log System Refactor - LogTape integration
- MongoDB 7.0 - Updated database support
- Enhanced MCP - Bidirectional MCP support
- Mobile Optimization - Improved mobile UX
January 2026:
- Workflow Demo Mode - Testing and demonstration features
- S3 Storage - Enhanced file storage options
- System Tool Categories - Organized tool management
February 2026 - Version 4.14.7:
- Context-Based Agent Mode - Beta for long task decomposition
- LLM Request Tracking - Configurable retention
- Knowledge Base Filtering - Filter by collectionIds
- Model Monitoring - Cache hit rate metrics
- Enhanced Chat Logs - Error filtering, user filtering
March 2026 - Version 4.14.9:
- AI Virtual Machine - New functionality
- AgentV2 Context Adaptation - Improved agent capabilities
- Workflow and Chat Bug Fixes - Stability improvements
- Release Date: March 22, 2026
| Date |
Milestone |
| 2023 |
Initial FastGPT release |
| 2024 Q1 |
Visual workflow builder |
| 2024 Q2 |
Hybrid search & reranking |
| 2024 Q3 |
MCP integration |
| 2024 Q4 |
Voice support |
| 2025 Q1 |
Agent mode planning |
| 2025 Q2 |
License update (Apache-2.0+) |
| 2025 Q3 |
LLM request tracking |
| 2025 Q4 |
Log system refactor |
| 2026 Q1 |
Workflow demo mode |
| Mar 2026 |
Version 4.14.9 (March 22, 2026) |
| 2026 |
27.5k+ GitHub stars |
Early Versions:
- Simple RAG pipeline
- Basic knowledge base
- MongoDB only
Current Versions:
- Hybrid search (vector + full-text)
- Reranking support
- MongoDB + PostgreSQL
- Visual workflow orchestration
- MCP integration
Evolution:
- Simple Q&A - Basic question-answering
- Workflow Nodes - Modular node system
- Visual Builder - Drag-and-drop interface
- Plugin System - Reusable plugins
- Agent Mode - Planning agents for complex tasks
Evolution:
- MongoDB - Primary database (conversations, knowledge base)
- PostgreSQL - Added for structured data
- Vector Search - Integrated vector database capabilities
Version 1.x API:
- Basic completions endpoint
- Simple knowledge base API
Current API:
- Full OpenAI compatibility
- Knowledge base CRUD
- Conversation CRUD
- Automated OpenAPI generation
- Webhook support
- Proprietary open-source license
- Some commercial use restrictions
- Community-focused
Apache-2.0 with Additional Conditions:
- ✅ Free for personal use
- ✅ Free for commercial use (as backend)
- ✅ Enterprise delivery allowed
- ❌ Multi-tenant SaaS restricted (need commercial license)
- ❌ Cannot remove LOGO/copyright (need commercial license)
Commercial License:
- Contact: dennis@sealos.io
- For SaaS offerings
- For white-label deployments
¶ Community and Ecosystem
Official Repository:
- labring/FastGPT
- 27.5k+ stars
- 7k+ forks
- 2,667 commits
- 168 contributors
- 223 releases
| Metric |
Value |
| GitHub Stars |
27.5k+ |
| Forks |
7k+ |
| Contributors |
168 |
| Releases |
223 |
Supported Integrations:
- LLM Providers: OpenAI, Anthropic, Qwen, DeepSeek
- Local Models: Ollama, llama.cpp
- Vector Databases: MongoDB with vector search
- MCP Servers: Model Context Protocol
- Cloud: Sealos Cloud, Kubernetes
FastGPT has significantly influenced the enterprise AI space:
- RAG Accessibility - Made RAG accessible to non-technical users
- Visual Orchestration - Popularized visual AI workflow building
- Enterprise Focus - Focused on enterprise deployment needs
- Open Source - Apache-2.0 license encourages adoption
- MCP Integration - Early adopter of Model Context Protocol
- 27.2k+ GitHub stars
- Widely adopted in enterprise deployments
- Featured in AI and enterprise software communities
- Recommended for RAG applications
- Used for customer support bots, knowledge management
- Latest Version: v4.14.7.2 (February 26, 2026)
- Release Cadence: Regular updates
- Active Development: Continuous feature additions
- Bug Tracking: Public GitHub issues
- Visual workflow orchestration
- Hybrid search & reranking
- MCP (Model Context Protocol) support
- Agent mode planning
- LLM request tracking
- Voice input/output
- Multi-format knowledge base
- OpenAI-compatible API
- Operations dashboard
- Template marketplace
Open-Source:
- Apache-2.0 with additional conditions
- Free for most use cases
- Community-driven development
Commercial License:
- Required for multi-tenant SaaS
- Required for white-label deployments
- Contact: dennis@sealos.io
Based on development patterns and public communications:
- Enhanced agent capabilities
- Improved workflow debugging
- Better MCP integration
- More template marketplace content
- Enhanced mobile experience
- Maintain Apache-2.0 license
- Expand enterprise features
- Enhanced collaboration tools
- Better multi-tenancy support
- Improved analytics and monitoring
- More LLM provider integrations
- Focus: Cloud-native infrastructure and AI platforms
- Products: FastGPT, Sealos Cloud
- Location: China (global operations)
- Mission: Simplify cloud-native and AI deployment
- Website: https://fastgpt.io
- GitHub: https://github.com/labring/FastGPT
- Documentation: https://doc.fastgpt.ai
- Commercial License: dennis@sealos.io
FastGPT represents an enterprise-focused approach to AI Agent development. Its evolution from simple RAG system to a platform with visual workflow orchestration, MCP support, and agent mode reflects the broader trajectory of enterprise AI adoption.
Key principles that guide FastGPT:
- Enterprise-Ready - Focused on production deployment
- Visual Orchestration - Easy workflow building
- Open Source - Apache-2.0 license (with conditions)
- RAG Excellence - Best-in-class RAG capabilities
- Flexibility - Support for multiple LLM providers
💼 Professional Services: Need expert help with your FastGPT deployment? We offer consulting, training, and support. Contact our team →