Skip to content

Building ETL Pipelines Without Code

Tags: guide, workflow Author: AgentBrain Team Published: 2026-02-10 Read time: 6 minutes

Source note: the canonical article body is in active drafting on agentbrain.sh and currently exists as a stub. The summary below captures the topic; this page will be updated with the full body when it lands.

What the Visual Workflow Editor Does

A drag-and-drop builder for multi-step data sync and transformation pipelines. Shipped in v0.9.0 (2026-02-15) with 10+ operators and cron scheduling.

Available Operators

  • Source operators: read from a configured data connector
  • Transform operators: filter, map, enrich, deduplicate
  • AI operators: agent call, classification, embedding generation
  • Sink operators: write to a Knowledge Vault collection, send to a channel, call a webhook
  • Control operators: conditional branch, parallel fan-out, human approval gate

Trigger Modes

  • Cron: on a schedule (every hour, daily at 03:00, etc.)
  • Webhook: triggered by external events
  • Manual: run on demand from the UI
  • Event: triggered by other workflows or agent actions

Typical Use Cases

PatternSource → Sink
Knowledge refreshNotion → Knowledge Vault collection
Daily reportPostgreSQL → Email summary
Trigger responseWebhook → Agent → Slack
Cross-system syncClickUp → ERP

How This Differs From Generic Workflow Tools

  • Native AgentBrain RBAC: workflows inherit the creator's permission scope
  • Audit-logged: every operator run is captured in the same audit log as agent actions
  • Agent-aware: agent operators are first-class and inherit the same identity and routing rules

Get Started

  1. Open Workflows → New in the AgentBrain console
  2. Drag operators onto the canvas
  3. Configure each operator (the right-hand panel shows the inputs)
  4. Wire them together
  5. Test with the run-once button
  6. Schedule via the trigger panel