Classify academic discussion with depth, CAPITAL and objectivity.
InsightGrade reads individual posts and full threads, classifies learning depth, generates a CAPITAL profile, performs a bias self-check, and returns everything in strict JSON — ready to analyse, compare and cite.
InsightGrade analyzes academic discussion posts and related instructional materials that may constitute education records under FERPA. To ensure compliance and maintain required institutional control, a signed Institutional Agreement must be in place before any institution activates or subscribes to a paid InsightGrade analysis plan. Please contact us at think@insightgrade.app to establish an agreement, or download the signable agreement.
InsightGrade is accountable for the service; Lovable provides contracted hosting and AI infrastructure. Our signed DPA prohibits model training on customer personal data and defines retention and deletion obligations. Read the data-protection summary and signed DPA.
One instrument, four layers of analysis
Learning depth classification
Every post is classified as deep, surface, or social learning based on the reasoning it actually contains.
CAPITAL profile analysis
Confidence, Amicability, Professionalism, Interactivity, Transparency, Adaptability and Lexicography — scored per post.
Bias self-check
A built-in audit confirms no demographic attributes influenced scoring — only content and reasoning.
Optional de-identification
Tick one box and student names are swapped for pseudonyms on your own device — for pasted text and for PDFs, whose text is extracted and cleaned in your browser so the original file is never uploaded.
Strict JSON output
Classifications, profiles, thread summaries and bias checks come back in one predictable machine-readable schema.
Communication analysis
The CAPITAL dimensions
CAPITAL is a multidimensional communication analysis framework used to evaluate discussion posts independently of learning type. Each dimension is scored or labeled strictly according to observable linguistic evidence in the post. CAPITAL dimensions are never inferred from author identity, tone stereotypes, or demographic assumptions.
- C — Confidence
Confidence
The degree of assertiveness, clarity, and decisiveness in the writer’s claims.
Scored from 0.0 to 1.0, with a one-sentence justification based solely on textual evidence.
Higher confidence appears as clear claims, minimal hedging, and decisive language. Lower confidence appears as hedging, uncertainty, self-doubt, or tentative phrasing.
- A — Amicability
Amicability
Warmth, friendliness, and interpersonal positivity expressed in the post.
Labeled low, moderate, or high.
- P — Professionalism
Professionalism
Formality, clarity, and academic appropriateness of language.
Labeled low, moderate, or high.
- I — Interactivity
Interactivity
Engagement with peers’ ideas, questions, or contributions.
Labeled low, moderate, or high.
- T — Transparency
Transparency
Explicitness of reasoning, clarity of thought process, and openness about uncertainty.
Labeled low, moderate, or high.
- A — Adaptability
Adaptability
Responsiveness to context, flexibility in perspective, and ability to integrate new information.
Labeled low, moderate, or high.
- L — Lexicography
Lexicography
Vocabulary richness, precision, and appropriateness for the academic context.
Labeled low, moderate, or high.
How it works
- 01
Submit
Paste a post or upload a discussion thread in plain text, PDF or Word.
- 02
Analyse
Depth is classified, CAPITAL dimensions scored, and a bias self-check run.
- 03
Export
Receive a polished executive report with per-post results and thread-level summaries.
Have a question?
Ask the InsightGrade Guide
Get instant answers about how analyses work, what the CAPITAL profile measures, which plan fits you, and how InsightGrade handles FERPA, student privacy and data protection.
InsightGrade Guide
Questions about how it works, plans, or FERPA
Ask anything about InsightGrade
Automated guidance, not legal advice. Don't share student information here. For anything else, email think@insightgrade.app.
Built-in disciplinary exemplars
Depth and CAPITAL scoring is calibrated against exemplar posts written for each field, so "deep" means the same thing in a nursing thread as it does in a philosophy seminar.
Social sciences & humanities
- Psychology
- Sociology
- Literature
- History
- Philosophy
STEM
- Biology
- Nursing
- Computer Science
Applied & professional
- Business
- Education
- Criminal Justice
Material from other fields is analysed against the closest exemplar set, and the analysis says which one was used.
What you get back
Per-post analysis
- Learning depth label: deep, surface or social
- CAPITAL scores with evidence from the post
- Confidence and reasoning for each classification
Thread-level summary
- Aggregated depth distribution across the thread
- Overall CAPITAL profile and trends
- Key observations and recommendations
Bias self-check
- Confirms no demographic attributes were used
- Documents the basis of every judgement
- Supports reproducible, defensible evaluation
Executive summary
- Polished, readable report for every run
- Easy to share with colleagues and committees
- Includes excerpts and scoring rationale
Start coding your discussions today
Upload a thread and get a full depth, CAPITAL and bias analysis in minutes.
Create your account