sirithreddy175-boop/interviewflowai

Purpose and Runtime

This is a TanStack Start/React TypeScript application for browsing candidates and running AI-assisted interview sessions. It is served through the TanStack Start server runtime, with Vite-based development/build tooling and Netlify deployment configuration in netlify.toml.

Primary startup and routing entry points:

• src/start.ts

• src/server.ts

• src/router.tsx

• src/routes/__root.tsx

Route and UI Layer

TanStack file routes define the user-facing flows:

• Landing/dashboard: src/routes/index.tsx

• Candidate browsing: src/routes/candidates.tsx

• Interview experience: src/routes/interview.tsx

• Interview backend endpoint: src/routes/api/interview.ts

The generated route manifest, src/routeTree.gen.ts, connects route files to the router. The application is a client-rendered React interface with server-capable API routes rather than a standalone REST service.

Candidate and Curriculum Domain Data

Candidate records and the interview curriculum are local domain data sources:

• src/data/candidates.ts

• src/data/curriculum.ts

• src/lib/interview/curriculum-service.ts

The candidate route and interview flow consume this data to select a candidate and determine interview topics/questions. curriculum-service.ts forms the boundary between static curriculum content and the interview engine.

Interview Session Orchestration

The central interview domain is implemented as server-side orchestration:

• src/lib/interview/engine.server.ts

• src/lib/interview/session-store.server.ts

• src/lib/interview/types.ts

The engine coordinates session initialization, interview progression, answer handling, and feedback-oriented state transitions. The session store separates mutable interview state from route/API handling. Shared types and constants constrain the data exchanged among the UI, API, engine, and AI provider:

• src/lib/interview/constants.ts

• src/lib/interview/types.ts

AI Integration Boundary

AI generation/evaluation is isolated behind:

• src/lib/interview/ai-service.server.ts

• src/routes/api/interview.ts

The API route accepts browser requests, invokes the server-only interview engine, and returns session/interview results. Keeping AI access in .server.ts modules prevents provider credentials and model calls from entering client bundles. Prompt-related project guidance is also present in AI_PROMPTS.md.

Client Interaction and Presentation

The browser-side service encapsulates calls to the interview API:

• src/services/interview-api.ts

Interview-specific presentation components render candidate context and resulting assessment:

• src/components/interview/CandidateCard.tsx

• src/components/interview/CandidateProfile.tsx

• src/components/interview/FeedbackReport.tsx

These components sit above the shared UI component library in src/components/ui/, keeping domain views distinct from reusable controls.

Persistence, Auth, and Error Boundaries

Supabase integration provides typed client/server access and auth request attachment/middleware:

• src/integrations/supabase/client.ts

• src/integrations/supabase/client.server.ts

• src/integrations/supabase/auth-middleware.ts

Database schema changes are represented in:

• supabase/migrations/20260808063713_570c7932-53f8-4464-add9-0ac4ed07b7af.sql

Error capture and error-page behavior are centralized outside feature code:

• src/lib/error-capture.ts

• src/lib/error-page.ts

• src/lib/lovable-error-reporting.ts