github.devotd.
NewRAG & retrieval

1panel-dev/maxkb v2.10.7-lts — Agent platform based on a document knowledge base

1panel-dev/maxkb is a Python-based open-source agent platform that builds knowledge bases through document collection and retrieval-based question answering, and composes workflows.

Repository1panel-dev/maxkb ↗

1panel-dev/maxkb is a Python-based open-source agent platform that builds knowledge bases through document collection and retrieval-based question answering, and composes workflows.

GitHub description: 🔥 MaxKB is an open-source platform for building enterprise-grade agents. 强大易用的开源企业级智能体平台。

stars
★ 23K
Forks
3.2K
Language
Python
License
GPL-3.0
Latest release
v2.10.7-lts
Read original source ↗View on GitHub ↗

1panel-dev/maxkb is an open-source platform for building AI agents that gather documents into knowledge bases and answer questions on the basis of documents found through search. It supports a RAG (retrieval-augmented generation) pipeline that feeds the retrieved content into the model to find the answer. The repository's latest release is v2.10.7-lts, written in Python and published under the GPL-3.0 license. However, the given materials do not list what changed in v2.10.7-lts.

The README describes MaxKB as a platform for building enterprise agents. It collects material through document uploads and online document crawling, and builds a searchable knowledge base with automatic chunking and vectorization (the processing that turns documents into numeric representations so that similar content can be found). On top of this come a workflow engine, a function library, and the use of MCP (Model Context Protocol) tools. It integrates with third-party business systems without coding, and it is stated to handle both public and private models and to support text, image, audio, and video.

That said, the verifiable range is limited. The feature list is based on the README description, and the materials contain no measurements showing search accuracy, response latency, or operating cost. Scores for novelty, technology, and quality are not given either, and the momentum score is low at 3.0 out of 10. The repository was created on September 14, 2023, but actual quality and operating cost are difficult to judge from this record alone.