Model ManagementManage models from OpenAI, DeepSeek, Qianwen, Zhipu, Ollama, and custom OpenAI-compatible APIs in one place.Configure models
Platform IntegrationConnect FastGPT, RAGFlow, and OpenAI- or Anthropic-compatible services. Dify and Coze adapters require additional provider options and credential support.Connect a platform
Multimodal & MediaExplore provider routing and APIs for images, speech, video, and predictions, including the current limits of attachments in ordinary chat.Explore media capabilities
Knowledge BaseFollow the RAG pipeline from document parsing and chunking to embeddings, vector storage, retrieval, reranking, and hybrid search.Set up retrieval
Tool ManagementUse Function Calling, built-in Java tools, local and remote MCP servers, and tool marketplaces.Configure tools
MCP ManagementLearn the protocol, connect LOCAL and REMOTE servers or ModelScope services, and verify agent tool calls with backend diagnostics.Connect MCP services
SkillsManage docx, pdf, and xlsx resources, with clear distinctions between chat selections and execution through Coding Harness.Understand skills
Context ManagementUnderstand prompts, history, and retrieved passages. Verify the 20-message window and extend persistence or long-term memory with LangChain4j examples.Explore context and memory
AgentsUse a LangChain4j Supervisor to coordinate search, SQL, charts, ECharts, and conversation, with knowledge retrieval and MCP tools.Build an agent
Workflow OrchestrationDesign and run model, retrieval, and branching workflows visually over SSE. Initialize optional nodes for mail, HTTP, and images as needed.Design a workflow