BullFlow
推量牛 BullFlow — 开源企业级 RAG 解决方案:从企业知识库、营销内容自动化生产到全域内容分发 | Open-source enterprise RAG solution: knowledge base → automated mark
What is BullFlow?
推量牛 BullFlow — 开源企业级 RAG 解决方案:从企业知识库、营销内容自动化生产到全域内容分发 | Open-source enterprise RAG solution: knowledge base → automated marketing content → omnichannel publishing. 知识库全链路本地开源模型,无需付费模型;AI Agent 创作图文 / 图片 / 视频,一键发布抖音、小红书、公众号、LinkedIn、YouTube 等 17 个平台。Commercial-grade, multi-tenant, billing & admin built in. MIT
Overview of BullFlow
What it is
BullFlow (推量牛) is an open-source enterprise-grade RAG (Retrieval-Augmented Generation) solution. It transforms enterprise knowledge bases (documents, images, audio/video) into automated marketing content and distributes it across multiple social media platforms. The entire RAG pipeline uses only open-source models running locally, with no paid model dependencies. It includes three complete systems: user frontend, operation admin backend, and a distribution gateway. The project is licensed under MIT and is free for commercial use.
Key features
- Enterprise knowledge base with structured archives (company, brand, product, customer, case) and industrial-grade RAG (OCR, semantic chunking, hybrid search with Milvus + Elasticsearch, reranking, Ragas evaluation).
- AI Agent (LangGraph ReAct) for content creation: articles with embedded images, multi-image posts, short videos, and voice narration.
- One-click distribution to 17 platforms (Douyin, Kuaishou, Xiaohongshu, WeChat Video, Bilibili, Zhihu, Weibo, Toutiao, Baijiahao, Duoduo Video, LinkedIn, Facebook, Instagram, Reddit, YouTube, X) with automatic format validation.
- Built-in commercial capabilities: multi-tenant, role-based access, token-based billing, subscription plans, payment (Alipay/WeChat/transfer), invoices, referral commissions, and conversion funnel analytics.
- Fully self-hosted: all components (vector DB, OCR, speech, models) can be deployed on-premises; models are packaged into Docker images with zero runtime downloads.
Use cases
- Brand social media matrix management: create once, publish to multiple platforms automatically with scheduled tasks.
- Cross-border e-commerce/overseas marketing: generate platform-specific content (LinkedIn, Facebook, Instagram, Reddit, X, YouTube) from the same product materials.
- ToB enterprise content marketing: upload whitepapers, case studies, product docs; AI writes industry articles with traceable citations.
- MCN/agency operations: multi-tenant isolation, each client's knowledge base and accounts are invisible to others; billing system supports reselling.
- Local business/store marketing: product images + one-line requirement → Xiaohongshu multi-image notes with copywriting.
- Enterprise knowledge Q&A: query documents, PPTs, Excel, audio recordings in natural language with source citations.
Getting started
See the repository README for installation and quick-start commands.
License and pricing
Licensed under MIT. Open source and free for commercial use, modification, and redistribution.
Based on the project's public README and repository metadata.