Skip to Content

Platform Capabilities Overview

Node orchestration is only the skeleton of the platform. Around building, running, securing, and collaborating on Agents, the platform provides a full set of supporting capabilities, each covered in its own article:

Capability Index

Core Capabilities (directly determine an Agent's ceiling)

#CapabilityOne-line Summary
01Knowledge BaseUpload private documents so the Agent answers based on your knowledge
02Skill SystemOn-demand capability extension packages with dynamic context management
03Workflow TriggersScheduled / Webhook auto-triggering for unattended Workflow runs
04GuardrailTwo-way compliance interception on input and output: PII masking, blocking of harmful content and financial violations
05Context & Memory ManagementShort-term memory, context passing, and auto-compression, so multi-turn conversations never lose track
06Planning & ExecutionTodo planning, Human in the Loop, Agent answer mode

Supporting Capabilities (make building and operations more efficient)

#CapabilityOne-line Summary
07Running, Debugging & ObservabilityThree run modes (node/full/production), GUI batch runs, Logs, statistics dashboard, and Langfuse tracing
08Security Verification & LocalizationTrading password verification, automatic Simplified/Traditional Chinese conversion
09Canvas & Collaboration EfficiencyCopying nodes across Agents, workflow locking, member permission switching, and more
10Model & Tool EcosystemSOTA model pool, account/community/P&L MCP tools, time zone optimization for market data tools
11Official Chatbot ExperienceConversation charts, voice input, long-term memory, and other official Chatbot (LongbridgeAI chat client) capabilities, sharing usage quota with the platform
12Disaster RecoveryCross-vendor model calls, model combinations, failure retries, exception handling

Recommendations

  • Before building your first Agent: no need to read everything — start with Quick Start
  • Agent needs private knowledge → 01 Knowledge Base; needs professional capability packages → 02 Skill
  • Workflow should run automatically → 03 Triggers; Agent is customer-facing → 04 Guardrail (must-read)
  • Multi-turn conversations forget earlier context / context grows too large → 05 Context & Memory
  • Complex tasks running out of control, need human oversight → 06 Planning & Execution

Entry points and features are subject to the current product version. When new capabilities are released, add or update the corresponding document in this directory and register it in the table above.