eltPulse

Getting started

eltPulse uses Clerk for authentication. After you sign up, we provision your workspace and you land on the dashboard. No Docker or gateway setup is required — managed execution is the default.

Fastest path: Quick start

  1. Open Quick start from the sidebar.
  2. Pick a destination (e.g. Snowflake, BigQuery, Postgres), then a source.
  3. Enter credentials (or link a saved connection) and click Create & run.

eltPulse generates production-ready dlt or Sling code, starts a run on managed workers, and streams telemetry to the Runs page.

Full builder: catalog, Pulse AI, or manual

  1. Open Pipelines.
  2. Choose Browse catalog, Pulse AI, or Manual create mode.
  3. Set a name (snake_case), source, and destination.
  4. Configure credentials inline or link saved connections. Use Guided mode for structured connectors or JSON for advanced config.
  5. Optional: add post-load dbt, configure run slices, or open the visual canvas after save.
  6. Submit — we store the definition and generate pipeline.py or replication.yaml, config.yaml, and eltpulse_workspace.yaml.

Visual canvas

The canvas is a visual graph editor for the same pipeline as the form builder. Drag source, transform, and destination nodes; wire edges; inspect dbt and post-transform settings on transform nodes. Pulse AI can patch the graph from chat — see Pulse AI.

Review generated files

From the pipeline table, open Code to copy artifacts. When GitHub is connected under Repositories, pipeline YAML can auto-push on save (Pro+; disable with ELTPULSE_AUTO_GIT_PUSH=false). Manual copy still works into eltpulse/pipelines/<name>/.

Credentials

Save reusable profiles on Connections — secrets are encrypted at rest. Pipelines link optional source/destination connection ids; generated YAML may include resolved connection names for runners. Gateways fetch decrypted profiles via GET /api/agent/connections.

Next

  • Concepts — pipelines, runs, connections, monitors, gateways.
  • Pulse AI — natural-language pipeline creation.
  • Run slices — backfills and incremental windows.
  • Connectors — catalog tiers and sync modes.
  • Schedules — cron on pipelines and dbt phases.
  • Gateway — when you need execution in your VPC.