Systems / integration
API integrations connect product actions, authenticated users, AI workflows, data storage, and reporting systems into one working product surface.
What this system does
Turns product actions into reliable backend requests with validation, access checks, and response states.
Connects SaaS services such as Clerk, Liveblocks, Trigger.dev, OpenRouter, Vercel Blob, and Postgres-backed persistence.
Supports event capture and workflow handoffs so product usage can become analytics evidence later.
How I use it
Ghost-AI uses authenticated project APIs, Liveblocks auth, AI design task routes, and secure spec download routes.
BudgetDB uses SQL files and workbook artifacts as source-to-output evidence rather than isolated screenshots.
The portfolio itself links code files, dashboards, documents, GitHub repos, and live deployments through reusable resource objects.
Related projects
Evidence or examples
Ghost-AI project overview defines project creation, collaborator access, Liveblocks room auth, AI generation, and spec download flows.
Ghost-AI architecture notes separate request handlers, background tasks, shared infrastructure, UI components, and database schema.
Portfolio project cards use functional links for GitHub, live deployments, SQL files, Python files, workbooks, and dashboards.
User action -> Clerk auth -> /api/projects route -> ownership check -> Prisma/Postgres metadata -> UI update.Prompt + roomId + projectId -> /api/ai/design -> Trigger.dev run -> TaskRun record -> run-scoped public token.Request spec -> verify project access -> verify spec belongs to project -> fetch Vercel Blob file -> return Markdown attachment.product_event(user_id, event_name, entity_id, occurred_at, metadata) -> funnel, retention, activation, and usage analysis.