AutomationCAN I VIBECODE
ServiceNow?
BUILD FIRST
- A non-sensitive or read-only version of the core experience
- A manual review step wherever correctness or trust matters
- A narrow companion tool that removes a real piece of work
LEAVE FOR LATER
- Regulated decisions, payments, or irreversible actions
- Claims of correctness, compliance, or uptime that the MVP cannot support
- Sensitive integrations until security, support, and auditability are ready
RESEARCH NOTES
What this tool actually does.
ServiceNow is an enterprise cloud platform (the 'Now Platform') for digital workflows and business automation. It provides no-code/low-code builders to automate multi-step processes across IT, HR, customer service, and other departments. Core workflow tools include Workflow Studio (a single design environment consolidating flows, playbooks, actions, decision tables, and integrations), Flow Designer (multi-step digital workflows across apps/systems), Decision Builder (no-code decision tables), Agentic Playbooks (no-code management of complex multi-agent workflows), Integration Hub (connect to modern-API systems), RPA Hub (robotic process automation for legacy systems), and Automation Engine (unified hyperautomation). The platform emphasizes a single data model connecting to 450+ systems (e.g., SAP, Salesforce), native AI/GenAI for building workflows via natural language, and AI agents tha
RESEARCHED BUILD PROMPT
Give this to your coding agent.
Build a focused open-source replacement for the core workflow-automation experience of ServiceNow (the Now Platform). The product is a self-hosted web app where a user designs, runs, and monitors multi-step business workflows without writing code.
Workflow: (1) A user opens a visual Workflow Studio-style canvas and drags steps (triggers, conditions/decision tables, actions, human approvals, integrations) into a flow. (2) They configure each step via forms (e.g., decision tables with if/then rules, API call actions, email/notification actions). (3) They publish the flow, which runs asynchronously; the app executes steps, calls external HTTP APIs, and records each run's state. (4) A run-history/monitoring view shows per-run status, logs, and errors, with the ability to retry or cancel a run.
User-facing requirements: visual flow editor with drag-and-drop nodes and connectors; node property panels; decision-table builder; run history list with status, timestamps, and logs; per-run detail view; simple user auth and per-user workspaces.
Technical requirements: backend (Node.js/Python) exposing a REST API for CRUD on flows and runs; a workflow execution engine that walks the graph, evaluates conditions, invokes outbound HTTP actions, and persists run state; SQLite/Postgres persistence; a web frontend (React) rendering the canvas (use a library like React Flow or a custom SVG/canvas graph) and the monitoring UI; background worker (queue) for async execution.
Implementation approach: keep the engine deterministic and testable; store flows as JSON; use a job queue (e.g., BullMQ or Celery) for runs; add webhook/HTTP action support with timeouts and retries.
Do NOT copy ServiceNow's brand, logo, proprietary UI, or any copyrighted assets; use an original name and clean, minimal UI. Do not attempt to replicate the full Now Platform (no RPA, no AI agents, no 450-system integration catalog, no enterprise governance/compliance suite).
Deliberately out of scope: AI/LLM features, RPA, multi-tenant enterprise security/SSO, mobile apps, and the full integration marketplace.
Include a README with setup instructions (env vars, DB migration, running worker + API + frontend) and important limitations (single-node execution, no distributed orchestration, no production-grade auth, no AI capabilities).