MaxKB
MaxKB (Max Knowledge Brain), from the team behind 1Panel, is an open-source platform for building enterprise-grade agents with knowledge bases at their core. It combines a RAG pipeline, where you upload documents or crawl online docs and the platform splits and vectorises them automatically, with an agentic workflow engine, a function library and MCP tool use for more complex business processes.
It suits organisations that want internal Q&A assistants, customer-service bots or knowledge-driven agents without building the stack themselves, especially teams in China that need private deployment with domestic models. It is model-agnostic, supporting private models such as DeepSeek, Llama and Qwen as well as public ones such as OpenAI, Claude, Gemini and MiniMax; it handles text, image, audio and video input and output; and assistants can be embedded into existing business systems with little or no code.
Getting started: run the Docker command `docker run -d --name=maxkb --restart=always -p 8080:8080 -v ~/.maxkb:/opt/maxkb 1panel/maxkb`, open http://your_server_ip:8080 and sign in with the default admin account. Change the default password immediately. An offline installation guide is available for environments where pulling Docker images fails.
Features
Why choose it
Trade-offs
Maintenance check
- Last commit
- 2026-11-10
- Commits, last 90 days
- 100+
- Releases, last 90 days
- 3
- Latest release
- v2.10.8-lts (2026-10-22)
- Open issues
- 25
- Recent issues with at least one comment
- 1/3
Oldest week on the left; the latest four weeks are orange. Source: GitHub commit activity.
Status rule: archived on GitHub = Archived; last commit within 90 days = Active; 91 to 365 days = Slowing; older = Stale. All figures come from the GitHub API on the date shown.
Most discussed open issues
- #7069 [Bug] 对接 vLLM 部署Qwen3.8-27B模型,开启思考模式后前端无法显示思考过程8 comments · 2026-09-16
- #6480 [Bug] Anthropic models: tool calls fail with "Input tag 'input_json_delta' ... does not match any of the expected tags" (streamed tool_use n2 comments · 2026-07-21
- #237 [FEATURE]希望能增加prompts模板2 comments · 2024-04-24
- #7809 [Bug] 上传excel长文件 上传的文档跟文档内容提取的顺序不一致1 comments · 2026-11-10
- #7719 [Bug] 建议 增加模型的容错处理1 comments · 2026-11-06
Issue titles are quoted from GitHub. Open the link to check whether a problem still applies to the version you use.
Who it fits
- You want an all-in-one knowledge-base and workflow agent platform that starts with a single Docker command and embeds easily into websites.
- You need broad support for domestic Chinese LLM providers alongside global models for internal enterprise Q&A.
- You cannot use GPL-3.0 licensed software in your commercial distribution; RAGFlow is Apache-2.0.
- Your primary bottleneck is extracting complex tables and layouts from messy scanned PDFs; RAGFlow's deep document parser goes further.
How it differs
MaxKB (GPL-3.0) is lighter to deploy via a single Docker command and geared towards rapid customer-service Q&A embedding; RAGFlow (Apache-2.0) requires a heavier 16 GB RAM stack but offers deeper layout/OCR parsing and visual chunk inspection.
MaxKB is a self-hosted enterprise knowledge-base and workflow platform for internal documents; GPT Researcher is an autonomous agent that searches the live web and local files to write long cited reports.
Compatibility
Quick start
Use cases
Alternatives
Featured in these lists
Comments
No comments yet. Be the first!