====== Flowise ======
**Flowise** is an open-source, low-code platform for building AI agents, chatbots, and LLM workflows through a visual drag-and-drop interface. Built on top of [[langchain|LangChain]], it allows developers and non-developers to compose agents, RAG pipelines, and [[multi_agent_systems|multi-agent systems]] by connecting [[modular|modular]] nodes on a visual canvas. Flowise significantly lowers the barrier to creating sophisticated AI workflows.

  * **Website:** [[https://flowiseai.com|flowiseai.com]]
  * **[[github|GitHub]]:** [[https://github.com/FlowiseAI/Flowise|github.com/FlowiseAI/Flowise]](([[https://github.com/FlowiseAI/Flowise|github.com/FlowiseAI/Flowise]]))
  * **License:** Apache 2.0

===== Visual Builder =====
The **Node Editor** is the primary interface where users drag and drop pre-built components from a Component Library onto a canvas. Components include:

  * **LLMs:** GPT-4o, [[claude|Claude]], Gemini, Llama, Mistral, and other models
  * **Vector Databases:** [[pinecone|Pinecone]], [[weaviate|Weaviate]], Chroma, Qdrant, [[milvus|Milvus]]
  * **Tools:** Web search APIs, CRM integrations, database connectors, email services
  * **Memory Systems:** Short-term conversation memory, long-term vector storage, hybrid options
  * **Document Loaders:** PDF, CSV, web scrapers, API connectors

Users configure inputs (chat, API calls, file uploads), processing (LLM chains and tools), and outputs (JSON, text, API payloads) with real-time testing, tracing, and debugging.

===== Three Builder Modes =====
**Chatflow:** Conversational pipelines with memory management, token tracking, caching, and guardrails. Supports [[short_term_memory|short-term memory]] for context, long-term vector storage for knowledge, and [[human_in_the_loop|human-in-the-loop]] approval steps.

**Agentflow:** [[multi_agent_systems|Multi-agent systems]] with supervisor-worker patterns. Supports native agent-to-agent communication, conflict resolution, dynamic role assignment, and complex orchestration patterns.

**Assistant:** Rapid prototyping mode for quickly creating AI assistants with minimal configuration.

===== Key Features =====
  * **Performance:** Handles 1,000 concurrent connections with 150ms overhead and 200-300MB memory footprint
  * **Deployment:** Local, Docker, Kubernetes, or serverless; production-ready API endpoints with authentication and rate limiting
  * **Embeddable Widgets:** Chat widgets for embedding in websites; API/SDK access for programmatic control
  * **Version Control:** Git-style versioning for flows
  * **Real-time Collaboration:** Multiple users can work on flows simultaneously
  * **Cost Optimization:** 40-60% LLM cost savings via intelligent caching, deduplication, and model switching
  * **Guardrails:** Content filtering, hallucination detection, and [[human_in_the_loop|human-in-the-loop]] checkpoints
  * **Self-hosting:** As low as $6-8/month vs. $500-5,000 for hosted alternatives

===== Integrations =====
Flowise connects with numerous external services via API endpoints and pre-built nodes:
  * LLM providers ([[openai|OpenAI]], [[anthropic|Anthropic]], Google, local models via [[ollama|Ollama]])
  * Vector databases and document stores
  * Third-party APIs (CRM, ticketing, analytics platforms)
  * Platforms like Bubble.io via REST API
  * Proxy support for hiding API keys from client applications

===== Use Cases =====
  * Customer support chatbots with RAG-powered knowledge bases
  * Research agents with multi-step web search and analysis
  * Internal copilots for enterprise documentation
  * Data analysis pipelines with database integration
  * Embedded AI experiences in existing web applications
===== Related Pages =====
  * [[langchain|LangChain]]
  * [[retrieval_augmented_generation|Retrieval-Augmented Generation]]
  * [[tool_integration_patterns|Tool Integration Patterns]]
  * [[multi_agent_systems|Multi-Agent Systems]]

===== See Also =====
  * [[langflow|Langflow]]
  * [[chatdev|ChatDev]]
  * [[langgraph|LangGraph]]
  * [[github_copilot_vs_windsurf|GitHub Copilot vs Windsurf]]
  * [[streamlit_ai|Streamlit for AI]]

===== References =====