AutomationCAN I VIBECODE
Make?
BUILD FIRST
- One specific customer type and one primary use case
- A vertical slice that completes the main job without manual workarounds
- Visible limits, review steps, and a fallback when an external dependency fails
LEAVE FOR LATER
- Broad third-party integrations
- High-scale collaboration, syncing, and permissions
- Automation that can cause harm without human review
RESEARCH NOTES
What this tool actually does.
Make (formerly Integromat) is a visual-first no-code/low-code automation platform for building workflows called 'scenarios' that connect 3,000+ pre-built apps, data sources, and AI models. Users design scenarios on a drag-and-drop canvas where modules (triggers and actions) are chained, with support for routers, iterators, aggregators, data stores, webhooks, HTTP/API requests, custom JavaScript/Python, and 300+ GenAI app modules. Additional capabilities include AI agents built via natural language/custom prompts, AI-assisted mapping of fields, an MCP server/client, Make Grid (an auto-generated map of all automations), an analytics/observability dashboard, flow control (conditionals, error handlers), data transformation functions, scheduling, role-based access control, and per-team operation limits. Pricing is operations-based (credit consumption per executed operation) with a free tier a
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a self-hosted web application that recreates the core workflow of Make's 'scenario' builder: visually connecting app modules into automations that execute on a schedule or via webhook. Do not copy Make's brand, name, logo, marketing site, or their exact UI styling — create an original look and name, but keep the distinctive Make-like workflow of canvas-based scenario editing with modul nodes, flow control branches, and execution history.
Workflow: (1) User signs in and creates a scenario with a name and description. (2) User drags modules from a palette onto a freeform canvas: trigger (Webhook, Schedule/Cron, Manual), action (HTTP Request, Email Send via SMTP, simple Data Transform), logic (Router/Condition, Iterator), and Utility (Data Store get/set, Text/JSON transforms). (3) User connects module outputs to inputs with visual edges, mapping fields via a JSONPath-like expression editor. (4) User clicks Run to execute instantly and then toggles scheduling/cron or exposes a public webhook URL for permanent activation. (5) Every execution writes an immutable log entry with input, output, duration, and status per module; a per-execution trace view shows each step's result.
User-facing requirements: responsive canvas with pan/zoom, draggable nodes, edge creation by drag from output port to input port, inline field mapping UI with JSONPath autocomplete, execution log list with drill-down traces, webhook URL copy button, cron expression input with next-run preview, and role-based permissions (admin vs viewer). Also provide a README covering setup (env vars, database migration, seed data), architecture, and known limitations.
Technical requirements: use React Flow (or similar DAG canvas library) for the editor; Node.js/TypeScript backend with Express or Fastify; Postgres for storage with a jsonb column for scenario definitions; bullMQ or similar for scheduled execution; lightweight execution engine that walks the graph topologically, resolves JSONPath references, handles retries/timeouts, and writes execution traces; webhook receiver endpoint that enqueues runs; JWT auth; Docker Compose for local dev.
Out of scope (do not build): the 3,000-app integration catalog, AI agents/LLM-native features, Make Grid analytics dashboard, MCP server/client, multi-tenant billing/operations quotas, public marketplace, and pixel-matching Make's design system. Keep integration connectors to HTTP/SMTP built-ins only. In the README, state these limitations explicitly and note this is an educational replacement, not affiliated with Make.com.
The README must include setup steps (docker compose up, migrate, seed sample scenario), environment variables, how to add a new custom module type, and a 'Limitations' section that clearly states the product does not implement Make.com's catalog, AI features, or cloud infrastructure.