エージェントとマルチエージェント

BISHENG

dataelement/bisheng

BISHENG is an open-source LLM application DevOps platform designed for next-generation enterprise AI applications, offering comprehensive features like GenAI workflow orchestration, RAG, Agent management, and enterprise-grade system controls.

★ 11.8Kスター
⑂ 1.9Kフォーク数
118未解決の問題
Python言語
Apache-2.0ライセンス
Q@project.QualityScore編集部スコア

プロジェクトのスクリーンショット

BISHENGのスクリーンショット BISHENGのスクリーンショット BISHENGのスクリーンショット

概要

BISHENG is an open LLM application DevOps platform focusing on enterprise scenarios. Named after the inventor of movable type printing, the platform aims to support the widespread implementation of intelligent applications. It provides an independent and comprehensive application orchestration framework that allows human-in-the-loop interventions, complex logic execution like loops and parallelism, and high-precision document parsing. It is utilized by industry-leading organizations and Fortune 500 companies.

主な機能

  • Lingsight general-purpose agent with expert-level taste via Agent Guidance Language (AGL)
  • Unique BISHENG Workflow with independent and comprehensive application orchestration
  • Human-in-the-loop intervention during workflow execution
  • Visual flowchart-based logic for loops, parallelism, and batch processing
  • High-precision document parsing (printed, handwritten, rare characters, tables, layouts, seals)
  • Enterprise-grade features: RBAC, SSO/LDAP, security review, vulnerability scanning, traffic control
  • Unified model management, Evaluation, SFT, and Dataset Management

要件、インストール、クイックスタート

git clone https://github.com/dataelement/bisheng.git && cd bisheng/docker

使用方法

Run `docker compose -f docker-compose.yml -p bisheng up -d` from the installation directory. Access http://IP:3001 in a browser to reach the login page. The first registered user becomes the system admin by default.

モデルの互換性とユースケース

Not stated in the repository metadata

ライセンスとリスクに関する注意事項

Licensed under Apache-2.0.

Editorial verification 2026-08-09: repository URL, owner, description, license and repository statistics were reviewed. License metadata: Apache-2.0. README was fetched for the channel draft; re-check repository dependencies, releases and model terms before production use.

リリースとメンテナンス

Not stated in the repository metadata

Firecrawl

firecrawl/firecrawl

★ 161.1KTypeScript

LangChain

langchain-ai/langchain

★ 143.6KPython

RAGFlow

infiniflow/ragflow

★ 86.7KGo