1 · Quickstart
Bootstrap a local NoETL cluster, run the spike e2e, and read the diagnostic telemetry — about 15 minutes.
2 · GKE Production Deploy
Provision a GKE cluster, deploy the NoETL stack with Vertex AI as the triage backend, configure Auth0 + ingress + TLS, and run the spike against the cloud cluster. About 1–2 hours.
3 · Self-Troubleshooting Playbook
Compose a parent playbook that dispatches a failing sub-playbook, gets a structured diagnosis attached automatically, and routes downstream behavior on the diagnosis category. About 30 minutes.
4 · Frontend Onboarding
Long-form walkthrough for building a frontend that talks to the NoETL gateway: Auth0 flow, session token storage, GraphQL executePlaybook calls, SSE for live updates, error handling, and production hardening. About 45 minutes.
5 · Add a new MCP backend
Advanced. Build a new MCP backend behind the JSON-RPC contract used by mcp/ollama and mcp/vertex-ai. Worked example: AWS Bedrock. About 2 hours.
6 · Widget Rendering
Build a playbook step that emits a render descriptor, view it in the NoETL terminal-style prompt, and verify that nested widget args survive event projection. About 30 minutes.
. Travel agent + widgets
Flagship demo — a natural-language travel agent built from playbooks alone. Shows NoETL DSL as the templating layer for AI providers and MCP tools, with widgets as JSON over the wire.
. Travel GUI walkthrough
A screenshot-led end-to-end walkthrough of the NoETL travel agent: catalog registration, prompt widgets, provider switching, MCP hops, travel canvas, refinement forms, and audit validation.
. Trip planner end-to-end
Capstone walkthrough for the Adiona/muno trip-planner app: Muno chat UI, NoETL itinerary agent, Firestore event sourcing, Duffel test orders, Google Places, Amadeus hotels, and calendar widgets.
.7. Business data via playbooks
Step-by-step implementation guide for the no-direct-database pattern: SPA reads and writes business data exclusively through playbook executions. Built on the Adiona/muno trip planner using Firestore as the worked example; adapts to Postgres, Dynamo, your-backend-of-choice.
9 - Internet to Postgres to GCS
Create, register, and run a NoETL playbook that downloads public data, stores it in Postgres, then exports it to Google Cloud Storage.