Thuto Research Studio · AI-Powered Postgraduate Audio Research

THE INTERFACE · CUSTOMIZATION PANEL

Four sections.
One perfect audio.

Not another generic AI tool. Research-specific controls aligned to your PG guidelines and rooted in South African languages.

Thuto Research Studio · v2.1

InkubaLM Online NLLB Ready

1

Academic Format Selection

Tailored to postgraduate research outputs

Research Deep Dive

Two AI "experts" unpack your thesis, connecting theory to methodology and SA context.

Ch.2 Lit Review Ch.3 Method

Executive Brief

Bite-sized 5-min summary of key findings for supervisors who need a quick overview.

Ch.4 Findings Reports

Critical Review

Expert reviewer simulates NRF/HEQC panel: rigorous, constructive feedback on your work.

Pre-submission NRF

Peer Debate

Two AI hosts debate methodological choices or theoretical frameworks to strengthen argument.

Ch.1 Problem Ch.5 Discussion

Viva Voce Prep

Mock oral defense: AI asks likely examiner questions based on your discipline and identified gaps. The ultimate pre-submission dry run.

Final Thesis Dissertation Oral Defense

2

Postgraduate Guidelines

Your unique differentiator — audio tailored to SA requirements

Honours Mini-Dissertation

10-15K words

5 chapters · Introductory research project

Masters Thesis

20-30K words UP/Wits/UCT

Institution-specific templates and formatting

PhD Dissertation

80-100K words Comprehensive

Full doctoral research with original contribution

NRF Research Proposal

20 pages max

Standard NRF format with priority alignment

Journal Article

SUNSAN/Sabinet

South African accredited journal formats

Institutional Template (Optional)

🏛️ University of Pretoria

🏛️ Wits University

🏛️ UCT

🏛️ Stellenbosch

🏛️ UNISA

Upload yours

What this does:

• Audio commentary references specific section requirements per institution
• Flags missing required elements (e.g., "Ethics clearance statement not detected")
• Suggests institutional phrasing (e.g., Wits uses "Research Design" not "Methodology")

3

Multilingual Audio Output

InkubaLM + NLLB for all 11 official SA languages

🇬🇧 English (Primary)

🇿🇦 isiZulu

🇿🇦 isiXhosa

🇿🇦 Afrikaans

🇿🇦 Sesotho

🇿🇦 Setswana

🇿🇦 Xitsonga

🇿🇦 Tshivenda

🇿🇦 siSwati

🇿🇦 Sepedi

🇿🇦 isiNdebele

Mixed/Bilingual

Voice Preference

Academic

Formal tone

Conversational

Supportive tone

Socratic

Tutoring tone

Glossary Override

Add discipline-specific terms for correct pronunciation: Ubuntu, Indaba, Ubuntu pedagogy, Lekgotla, Botho.

4

Focus & Scope Customization

Research-specific prompts that evolve with your chapter

🎯 What should the AI hosts focus on?

Pre-loaded with Chapter 2 · Literature Review prompts

Suggested Prompts (click to add):

✓ Check alignment with NRF priorities 📚 Identify missing key theorists ⚙️ Assess methodology justification 🛡️ Review ethical considerations 🌍 Evaluate contribution to SA field 🔍 Check plagiarism & citation style

📏 Length

Short

3-5 min

Default

8-12 min

Long

15-20 min

🔊 Output Format

MP3

In-App

Playlist

LIVE PREVIEW

Your Audio Summary

Research Deep Dive · UP Masters · isiZulu

Chapter 3 · Methodology

Mixed methods · 28 pages

02:14 12:34

Format Deep Dive

Guideline UP Masters

Language isiZulu

Voice Academic

Length 12 min

Quality 44.1kHz · Stereo

AI Pre-Analysis

✓ 4 missing citations detected
✓ Population table needed (UP Ch.3)
⚠ Ethics statement flagged

INTEGRATION · LOOP OF EXCELLENCE

From audio to approval.

How Thuto Research Studio connects seamlessly into the supervision workflow.

1

Student generates Audio Overview

Chapter 3 (Methodology) · Critical Review format

2

Selects configuration

Masters Thesis (UP) · isiZulu · Academic voice

3

AI produces 12-min isiZulu critique

Audio + transcript + action items

4

Student listens, identifies weaknesses

Better retention than reading 40 pages

5

Submit via SuperviseAI

Draft + audio summary attached automatically

6

Supervisor plays audio summary

12 minutes vs 40 pages · 50% time saved

7

Professor records voice note feedback

Added to "Augmented Feedback" dossier

8

Student receives complete dossier

Audio feedback + annotated doc + action items

For Students

Listen to your work being critiqued — audio retention is 4× higher than reading, especially for students with reading difficulties or low bandwidth.

For Supervisors

Quick audio overview before deep reading. 50% time saved per chapter review. Voice note feedback is faster than typing.

For Accessibility

Audio-first for students with dyslexia, visual impairments, or low-bandwidth environments. 11 SA languages break the English barrier.

The loop that makes supervision human again.

Every audio generated becomes part of a growing dossier — a living record of your research journey, your supervisor's voice, and your intellectual evolution.

Postgraduate students using Thuto Research Studio

BUILT FOR SA CONTEXT

Your grandmother could understand your PhD.

With bilingual output, you can share a summary of your research in your home language. Perfect for community engagement reports, NRF impact statements, and explaining your work to the people who matter most.

Mixed-language hosts — one speaks English, another isiZulu. Perfect for community presentations.

Discipline glossaries — custom pronunciation for Ubuntu, Indaba, Lekgotla, Botho, and field-specific terms.

Low-bandwidth MP3 — works on 3G, downloadable for offline listening in rural areas.

COMMERCIALIZATION · AUDIO STUDIO REVENUE STREAMS

Priced for students.
Scaled for institutions.

Audio Add-On

For individual students

R29/50min

≈ R1,450/year (50 audios)

All 5 formats
All SA languages
MP3 download
Transcript included

Department License

Postgraduate offices

R15-45K/yr

Unlimited · R2.25M potential

Unlimited audio
Custom guidelines
Analytics dashboard
SSO integration
Dedicated support

Viva Voce Pack

Final-year PhDs

R99 once

R495K potential (5,000 PhDs)

Mock oral defense
Discipline-specific Qs
3 practice sessions
Examiner persona AI

Multilingual Export

Community engagement

R49/trans

R245K potential (5,000 exports)

Translate to any SA lang
Bilingual hosts
Community summaries
NRF impact reports

TECHNICAL IMPLEMENTATION · FOR MPHO

Under the hood.

Built on South African soil, for South African scholars.

ARCHITECTURE OVERVIEW

Frontend (React/Bubble)

Audio Studio UI · 4 sections

Backend API (FastAPI)

REST + WebSocket for streaming

🤖 AI PIPELINE

1. Document Parser → Extract text

2. Guideline Matcher → UP/Wits/UCT

3. Language Router → InkubaLM/NLLB

4. Format Selector → Deep Dive/Critique

5. Audio Generator → TTS (SA voices)

6. Output → MP3/WAV + Transcript

Storage · AWS S3 Cape Town

POPIA compliant · SA data residency

Delivery · CDN for streaming

Low-latency across SA

Key Endpoints

POST /api/v1/audio/generate

GET /api/v1/audio/languages

GET /api/v1/audio/guidelines

GET /api/v1/audio/voices

WS /api/v1/audio/stream

TTS Integration

ElevenLabs

English (SA accent) · Excellent

Microsoft Azure

11 SA languages · Good

InkubaLM + Custom TTS

All 11 official · Self-hosted

Recommended: Hybrid

Azure for SA langs, ElevenLabs for English

COMMON QUESTIONS

Frequently asked.

How accurate is the SA language output?

InkubaLM handles academic terminology with high accuracy for isiZulu, isiXhosa, and Afrikaans. For lesser-resourced languages (Tshivenda, Xitsonga), we use a hybrid approach: NLLB for structure + community-reviewed glossaries for terminology. All outputs include a confidence score and allow manual correction before final audio generation.

Does this replace my supervisor?

Absolutely not. Thuto Research Studio augments supervision. It prepares you for supervision sessions, flags issues before your supervisor sees them, and lets supervisors focus on high-value feedback rather than basic structural review. The Loop of Excellence ensures every audio becomes part of the supervisor-student dialogue.

How do you handle research confidentiality?

All documents are encrypted at rest (AES-256) and in transit (TLS 1.3). Data resides in AWS S3 Cape Town region for POPIA compliance. Documents are automatically deleted from our pipeline after audio generation unless you opt into persistent storage. We never train our models on your unpublished research.

Can I use it for my NRF proposal?

Yes — we have a dedicated NRF Research Proposal format that aligns with current NRF priority areas, checks page limits (20 pages max), validates budget tables, and cross-references your work against national development priorities. The Critical Review format is especially powerful for pre-submission polishing.

What if my university isn't listed?

Upload your institution's PG guidelines (PDF) and our system will parse and index them within 24 hours. Department license holders get priority indexing. We currently support UP, Wits, UCT, Stellenbosch, UNISA, and are actively onboarding UJ, NWU, UKZN, and Rhodes.

LAUNCH YOUR FIRST AUDIO TODAY

Your thesis deserves
to be heard.

Join the postgraduates across South Africa transforming silent research into spoken scholarship.

3 free audio generations · No credit card required

Hear your thesis think itself aloud.

Four sections.One perfect audio.