QuickFlows
Set up ingestion and ETL with source discovery, incremental sync and managed connections.
The durable execution plane
Build and run pipelines, workflows and agents.
Build with a visual canvas, declarative DAGs, SQL, Python or code-first agents. VegaFlow gives every mode the same durable state, governed context and operational history.
01
discover_catalog succeeded 00:0402
public.customers succeeded 00:3803
public.orders running 08:1204
public.order_items queued --:--05
06
cursor updated_at = 2026-09-03T00:44:12Z07
environment production / me-centralSTATUSREADYREGIONME-CENTRAL
VegaFlow / role in the platform
Build visually or in code. VegaFlow keeps track of versions, execution state, retries and approvals, so you can inspect a run or resume it after an interruption.
Capabilities
What you can build with VegaFlow
Declarative DAGs
Build versioned asset pipelines with explicit dependencies, schedules and observable execution.
SQL pipelines
Give analytics engineers a direct path from governed SQL to reliable production workflows.
Durable flows
Run long-lived, retryable processes with recoverable state and human-in-the-loop decisions.
Native AI agents
Build typed, stateful agents with tools, memory, guardrails and live organizational context.
Visual agents
Compose triggers, tools, decisions and human review on a canvas backed by the same runtime.
How it works
Follow the data and execution.
01
Trigger + plan
Resolve the version, policy, dependencies and organizational context for this run.
02
Execute
Run pipelines, services and agents on isolated compute with durable state.
03
Recover
Retry safely, resume long-running work and bring a person into the loop when needed.
04
Record
Record outputs, decisions and lineage in the shared graph.
VegaFlow / start with one outcome