FlowConnect AI

FlowConnect AI is a Next.js SaaS scaffold for workflow automation across app connectors, AI actions, hosted forms, REST APIs, transformation maps, usage billing, and a workflow template marketplace.

Run Locally

npm install
npm run dev

Open http://localhost:3000.

Database

Create .env.local from .env.example and set your private database credentials:

cp .env.example .env.local

Deploy the MySQL schema after the database environment variables are set:

npm run db:deploy

Check database connectivity from the app:

GET /api/health/db

Connection credentials are encrypted with a server-only encryption key. Set all secrets outside source control before saving real connections.

AI actions use server-only OpenAI environment variables. Keep the real API key only in .env.local or your deployment provider secrets.

Implemented Feature Areas

• Connector SDK with typed metadata, auth, scopes, triggers, actions, connection testing, token refresh, rate limits, and error handling.
• Starter SDK connectors for ServiceNow, GitHub, Google Sheets, YouTube, and Wix.
• Usage and billing engine with Starter, Pro, and Enterprise plan definitions.
• Usage dashboard for current plan, monthly usage, remaining workflow runs, and upgrade CTA.
• Workflow marketplace with browse, search, app/category filters, install/clone, ratings, official templates, and five starter templates.
• User signup, login, logout, profile, and HTTP-only session cookies.
• AI workflow proposal generation for connection-specific async API flows.
• AI form field generation for intake forms.
• API routes for connector registry, connector health tests, token refresh, usage tracking, template install/rating, and workflow draft/run creation.
• MySQL schema for users, connectors, workflows, runs, logs, usage events, usage rollups, templates, and template ratings.

Key Files

• lib/connector-sdk.ts: Connector contract and starter connector registry.
• lib/connector-runtime.ts: Stored connection runtime helpers for decrypting credentials, testing connections, and refreshing tokens.
• lib/usage-billing.ts: Pricing tiers, usage metrics, empty usage snapshot, and summary calculations.
• lib/usage-events.ts: Usage event creation, plan limit checks, and projected usage rollups.
• lib/marketplace.ts: Template definitions, search/filter helper, and clone helper.
• lib/workflow-engine.ts: Workflow validation and run record creation.
• lib/db.ts: MySQL pool configured from environment variables.
• lib/connection-secrets.ts: AES-GCM encryption for saved connection credentials.
• lib/auth.ts: User signup/login/session helpers.
• lib/openai-config.ts: Server-only OpenAI key/model config helper.
• lib/ai-workflow-proposal.ts: OpenAI-backed workflow proposal generation.
• lib/ai-form-proposal.ts: OpenAI-backed intake form proposal generation.
• sql/schema.sql: MySQL schema for the new platform features.

Adding a Connector

Add a new ConnectorDefinition entry in lib/connector-sdk.ts with:

• id
• appName
• appIcon
• category
• authType
• requiredScopes
• availableTriggers
• availableActions
• rateLimitRules
• errorHandlingRules
• testConnection
• refreshToken

The core workflow engine should call connector definitions through the registry instead of hard-coding app-specific logic.

Production Notes

• Store connection encryption secrets outside source control and plan key rotation before production launch.
• Persist usage events and monthly rollups in MySQL using sql/schema.sql.
• Enqueue workflow runs into BullMQ after createWorkflowRunRecord.
• Recalculate template ratings from template_ratings rather than trusting client-submitted aggregates.
• Enforce usage limits before workflow activation, connection creation, API calls, AI actions, form submissions, and storage writes.