Next.js · MongoDB · Gemini 2.5 Flash

Turning student
self-reflection
into structured care.

A digital bridge between students and counselors — quietly transforming free-form check-ins into clean, structured JSON insight, so counselors spend their time on people, not paperwork.

Explore the modules Why it was built
2dedicated portals
1AI analysis engine
JSONonly — no file uploads
The story

Why it was built

In modern university environments, a critical disconnect often exists between student mental health tracking and the actual bandwidth of academic counselors. Students continuously generate valuable self-assessment data, but counselors are frequently overwhelmed by the sheer volume of unstructured text they must review.

The Counseling Reports Portal was engineered to solve exactly that bottleneck — eliminating manual document parsing entirely. Instead of file uploads and messy formats, the platform runs on clean, structured JSON from end to end, so counselors can focus on human intervention instead of document triage.

Unstructured to structured

Free-form student input becomes consistent, query-able data the moment it's submitted.

Designed around trust

The interface is built so students feel heard, not analyzed — the technology stays invisible.

Time back to counselors

Less time parsing documents means more time for the conversations that actually help.

How it flows

From check-in to actionable insight

One synchronous pipeline, no intermediate files.

Student check-in

A short self-reflection submitted as plain text

Gemini 2.5 Flash

Synchronous deep-level analysis of the input

Structured JSON

Profile, risk and findings as enforced fields

Counselor dashboard

Ready for review and follow-up action

Two experiences, one system

Built differently for each side

For the Student — an empathetic interface

Frictionless, low-pressure, and deliberately calm.

  • Dashboard shows exactly when the next report is due
  • Plain-text submission — no forms, no file uploads
  • Success and status messages never mention "AI analysis"
  • Designed to encourage honest, unfiltered self-assessment

For the Counselor — the command center

Macro visibility with the detail to act when it matters.

  • Total assigned students & period-wise submission tracking
  • Capacity tools to set counseling ranges and specialties
  • Instant structured insight — profile, risk, recommended actions
  • 3–5 concise findings per report, ready for follow-up
Inside the platform

Nine modules, each one interactive

All sample data below is illustrative — generated for demonstration only.

Submission Tracking

A dedicated student dashboard that removes the guesswork of when something is due.

  • Tracks next report due date across the term
  • Shows progress across required periodic check-ins
  • Gentle reminders, no pressure-driven language

Sample student view

2 of 3check-ins this semester
06days left
Period 2current period
On trackstatus
  • Period 1 — Foundations check-inSubmitted Aug 14
  • Period 2 — Mid-term check-inDue Oct 02
  • Period 3 — End-of-term check-inOpens Nov 20

Anxiety-Free UX

The same event, described two different ways depending on who's looking — by design.

  • Student-facing copy never references AI or analysis
  • Internal/staff logs remain fully transparent about the pipeline
  • Builds trust so students self-report honestly

Same event, two audiences

1. Choose a scenario

2. Choose an audience

"Thanks for sharing — your check-in has been saved."

Counselor Command Center

A macro view of an entire caseload, refreshed as submissions arrive.

  • Total assigned students & overall submission counts
  • Period-wise tracking across the academic term
  • Surfaces who hasn't submitted yet

Sample dashboard

Assigned students86
Submitted this period71
Completion rate82%

Pending submissions — Period 2

StudentLast activeStatus
STU-22045 days agoReminder sent
STU-239111 days agoOverdue
STU-24562 days agoPending

Capacity Management

Counselors define their own bandwidth instead of being assigned an arbitrary caseload.

  • Adjustable counseling range / max caseload
  • Specialty tagging for better student-counselor matching
  • Profile stays current as availability changes

Counselor profile settings

Dr. A. MercerCounselor · since Jan 2023
52Currently assigned
8Open slots
87%Capacity used

Specialties

Academic stress Career guidance Anxiety support Peer conflict Time management

AI Analysis Engine

Gemini 2.5 Flash processes each submission synchronously and returns enforced JSON — never loose text.

  • Synchronous, deep-level text analysis
  • Strictly enforced JSON output schema
  • No multipart file handling anywhere in the pipeline

Run the pipeline

  1. Receive submission text
  2. Send to Gemini 2.5 Flash
  3. Validate against JSON schema
  4. Store structured record
{ "profile": { "dominantTrait": "—", "learningStyle": "—" }, "risk": { "level": "—", "score": 0 }, "findings": [] }
—Processing time
6Schema fields enforced
0Malformed responses

Psychological Profile

One layer of the structured output — a general behavioural snapshot, not a diagnosis.

  • Dominant trait & learning style
  • Motivation level scoring
  • Feeds counselor context before any conversation

Sample profile output

Record STU-2291-P2Generated from Period 2 check-in
Processed
Motivation
Engagement
Consistency
Dominant trait: ConscientiousLearning style: VisualUpdated 2 days ago

Risk Prediction

A general indicator that helps a counselor prioritize their day — always reviewed by a human, never acted on alone.

  • Calculated risk score & categorical level
  • General contributing-factor tags
  • Always paired with a human follow-up step

Sample risk summary — click a level

Risk score: 18 / 100 Routine monitoring — no immediate follow-up required.
Low

General contributing factors

Academic workloadSleep pattern

Flagged this period — caseload view

StudentRisk scoreLevel
STU-210424Low
STU-239158Medium
STU-221883High

Actionable Reports

The output a counselor actually reads — short, specific, and oriented toward next steps.

  • 3–5 concise counselor findings per report
  • Priority interventions ranked by urgency
  • Designed to be skimmed in under a minute

Click to generate a sample report

Counselor findings

  • Steady engagement with coursework this period
  • Mild increase in reported workload stress
  • Recommend a brief, informal check-in next week

Priority interventions

01Schedule a brief check-in within 7 daysLow
02Share academic workload resourcesLow

Tech & Data Flow

A pure JSON architecture from submission to dashboard — no legacy file-upload layer to maintain.

  • Next.js & React across both portals
  • MongoDB / Mongoose for flexible report schemas
  • @google/generative-ai for sanitized, parsed JSON output

Stack at a glance

Next.js & ReactStudent and counselor portals
MongoDB / MongooseFlexible, scalable report schemas
Gemini 2.5 FlashStructured JSON generation
Enforced JSON schemaNo multipart uploads, no loose text
0File uploads in pipeline
3Structured output sections
2Portals served from one schema
Technical foundation

Built to stay fast and predictable

Framework

Next.jsReact

Database

MongoDBMongoose

AI Integration

Google Generative AIGemini 2.5 Flash

Data Flow

Pure JSONNo file uploads