Documentation

Agents

View as Markdown

This section contains all documentation related to Agents in Clear Ideas.

Available Guides

  • Agents Overview - Use Clear Ideas Agents to run repeatable, document-grounded AI processes over approved Site content with reviewable runs, generated files, schedules, triggers, and evidence.
  • Agent Runs - Run, monitor, review, archive, duplicate, and export governed Agent runs in Clear Ideas.
  • Agent Builder - Use Build From Document to create an Agent from a reference file and approved Site context.
  • Agent Designer - Use Describe With AI to create production-ready Agents from a natural language description.
  • Triggers and Webhooks - Trigger Agents from webhooks, uploads, schedules, and other Agents with authenticated payload mapping and strict validation options.
  • Schedules - Use the Agent schedule calendar to review upcoming runs, recent runs, and recurring execution patterns.
  • Benchmarks - Evaluate Agent output quality, track improvement, and use benchmark feedback to refine repeatable AI processes.
  • Agent Connections - Connect Agents to Clear Ideas Sites and approved external MCP servers with controlled authentication, health checks, aliases, and evidence-aware source metadata.
  • Source Connections and Tool Policy - Configure the approved sources, required connections, allowed tools, and evidence settings that govern Agent runs.
  • Controlled Tools and Egress - Understand how Clear Ideas controls Agent tools, outbound calls, web access, generated files, and evidence for sensitive AI execution.
  • MCP Agent Authoring - Build and run Clear Ideas Agents from compatible MCP clients such as Codex, Claude, or Cursor while keeping context governed.
  • Agent Editor and Step Types - Use the Agent editor to configure sources, steps, models, outputs, web access, approvals, sub-agents, webhooks, and code steps.
  • Variables and Variable Sets - Define Agent variables, reuse variable sets, map trigger payloads, and keep scheduled or repeatable Agent runs consistent.
  • Trigger Mapping - Map webhook, upload, and sub-agent trigger payloads into Agent variables with required fields, default Sites, and strict validation.