Project Lab — the full guide

1. What Project Lab is

Project Lab is the part of CampusTutor built for one thing: helping you finish your academic project — your Final Year Project, seminar paper, mini project, internship report, thesis, or dissertation. It is not a generic AI writer. Every tool inside it is shaped around the way Nigerian university, polytechnic, and college-of-education students actually submit work.

You upload your real research materials (lecture PDFs, journal articles, past projects, scanned textbook pages, web links, your own notes). The AI reads them and writes chapters grounded in those sources — every factual claim cited back to a real reference. You edit, run quality checks, practise your defense, and export submission-ready DOCX, PDF, or LaTeX.

2. Plans & credits

Project Lab is a monthly or yearly subscription. Each tier comes with a monthly allowance of AI credits— the same credits power every AI action across the workspace. Unused credits roll over up to a cap so a slow week doesn't go to waste.

Defense Sprint (₦12,000, one-time) — when your viva is within 30 days we surface this option automatically. It unlocks 30 days of Postgrad Pro features plus 500 extra credits, then auto-cancels unless you choose to stay on Postgrad Pro.

Top-ups — running low mid-month? Top-up packs from ₦2,000 add credits that never expire: Mini (25 credits), Standard (70 credits), Bulk (200 credits, best rate).

What each action costs

One credit is roughly one user-facing AI call. Generating a fresh chapter section is the heaviest action; lightweight things like fact-checking, autocomplete, and file imports are free or near-free.

3. Starting a project

1. From the dashboard sidebar, click My Projects to open the Project Lab home.
2. Click New Project. If you don't see the button, choose View Plans & Pricing first — Project Lab needs an active subscription or a Defense Sprint.
3. Pick your project type: Final Year Project, Seminar Paper, Mini Project, Internship Report, Thesis, or Dissertation. Each one comes with a chapter scaffold tuned for that format.
4. Give the project a working title and a one-paragraph abstract describing what it's about. The abstract is the AI's anchor — clearer abstract, sharper chapters.
5. Add the department and level so the AI calibrates language for your audience (BSc vs MSc vs PhD reads very differently).
6. Click Create. You'll land inside the project workspace.

4. Workspace layout

The project workspace has three panels:

• Left — Chapter Sidebar. Your project's table of contents. Each chapter is a parent; sections live underneath. Click any section to load it in the editor. Status badges (Pending, Draft, Accepted, Edited, Locked) tell you what stage each one is at.
• Centre — Section Editor. The chapter you're currently writing. You can let the AI generate from scratch, type yourself, or paste content in. Autocomplete suggestions appear as you type — press Ctrl+J to accept.
• Right — Tool Panel. Ten tabs across the top: AI, Sources, Lit, Method, Topic, Diagrams, Refs, Plagiarism, Defensible, Stats. Each is a different superpower for the section you're currently editing.

Both side panels collapse and resize. Use the chevron buttons in the top bar to hide the sidebar or the tool panel when you want more room for the editor. Drag the divider between them to resize.

The top bar also shows a ← My Projects link on the left — click it any time to return to the project list. The current project title sits next to it as a breadcrumb so you always know where you are.

5. Writing chapters

Every section moves through a status lifecycle: Pending → Generating → Draft → Accepted → Edited → Locked. You can write entirely manually if you prefer — the AI is opt-in per section.

Letting the AI write

1. Open a Pending section.
2. Click Generate. The AI reads your project metadata, your uploaded sources, and the chapter outline, then writes a draft (7 credits, Claude Sonnet).
3. Review the draft. If it's broadly right, click Accept — status becomes Accepted. Now you can edit freely without re-generating.
4. To revise, open the AI tab on the right and use Improve, Expand, or Proofread. Each uses Haiku and costs 1–1.5 credits.

The AI tab — your section coach

The AI panel offers four sub-tools that operate on the current section:

• Improve — paste feedback ("make it more concise", "add more citations", "tighten the transition to the next section") and the AI rewrites with that guidance baked in.
• Proofread — checks grammar, academic tone, formality, clarity, and transitions. Flags hedging, informal language, and weak verbs.
• Expand — adds depth where the draft is thin. Optional focus hint lets you point at a specific aspect.
• Chat — ask anything about your project. The chat sees your sources, level, and current section, so it's context-aware.

Version history

Every generated or edited draft is automatically versioned. If a regeneration produced something worse than what you had, roll back from the section's history menu — nothing is permanently lost.

6. Sources & citations

The Sources tab (Research Library) is the foundation. Whatever you upload here becomes the factual evidence base the AI draws from when it writes. Anything not uploaded won't be cited.

What you can upload

• PDFs (journal articles, lecture notes, past projects)
• Word documents (.docx)
• Plain text files
• Images and scanned pages (.jpg, .png, .webp, .tiff) — OCR runs automatically, free
• Web articles via URL (paste a link, choose Web Article / Journal / Statistical Output / Past Project, click Add)
• BibTeX (.bib) or RIS (.ris) reference files from Zotero, Mendeley, or EndNote — metadata pre-fills instantly, free

Max file size is 20 MB per upload. Each uploaded file gets a quality tier(T1 = peer-reviewed Q1 journal, down to T5 = blog/website) so the AI knows which sources to lean on for serious claims.

How citations work

The AI cites as it writes — every factual claim gets an inline marker:

• APA / Harvard styles use (Author, year) format
• IEEE / Vancouver styles use numbered markers [1]
• Each marker traces back to the exact uploaded source — title, author, year, DOI, and the passage the AI drew from. Nothing is fabricated.
• Six styles supported: APA, MLA, Chicago, Harvard, IEEE, Vancouver. Switch with one toggle in the Refs tab — the entire list reformats instantly.

Auto-suggested sources

Open the Suggested Sources section at the bottom of the Sources tab. Project Lab searches OpenAlex— the world's largest open-access academic database — for papers relevant to your topic and surfaces them automatically. Add any to your library in one click with citation metadata included.

7. References & the Refs tab

The Refstab is where your full reference list lives. It builds itself automatically from every source you've uploaded — alphabetised, formatted, ready to paste at the end of your project.

• Toggle citation style at the top — the whole list reformats live.
• Click any reference to see (and fix) its citation metadata.
• Unverified or incomplete references show a warning icon — the Library Health bar at the top of Sources tells you how many need attention.
• The Auto-tag button asks the AI to classify each source into quality tiers T1–T5 in one pass.

8. Literature Engine (Postgrad Pro)

For dissertations and theses, the Lit tab gives you four power tools:

• Theme Map (3 credits) — clusters your sources into themes using semantic similarity. See at a glance how your literature breaks down.
• Gap Matrix — shows where the existing literature is thin. Your thesis sits in those gaps.
• Synthesis (1.5 cr per theme) — writes a synthesised literature review per theme that you can paste into Chapter 2.
• Theoretical Lineage Tree (2.5 credits) — visualises which theories your work descends from. Defence-ready.

9. Methodology Advisor (Postgrad Pro)

The Method tab walks you through five methodology decisions:

1. Paradigm wizard (0.8 cr) — picks positivist / interpretivist / pragmatist / critical realist based on your RQs.
2. Strategy (1.2 cr) — quant survey, qual interview, mixed-methods, case study, etc.
3. Instrument draft (2 cr) — first draft of your questionnaire, interview guide, or observation schedule.
4. Counter-design memo (2.5 cr) — the AI plays devil's advocate: "here's the rival design your examiner will suggest, and here's why you chose yours instead." Crucial defence prep.
5. Ethics statement (1 cr) — drafts the ethics paragraph for your methodology chapter.

10. Topic Engine

The Topictab helps you sharpen or replace a weak topic before you've invested in writing. Three tools:

• Kill-list check (0.1 cr) — searches a database of overused / supervisor-blacklisted topics. If yours appears, change it now, not after Chapter 3.
• External evidence (0.3 cr) — pulls Crossref & Semantic Scholar results so you can see how saturated your topic already is.
• Supervisor preview (0.4 cr) — drafts the kinds of objections a supervisor would raise. Pre-empt them before submission.

11. Diagrams

The Diagrams tab generates eight diagram types (1 credit each, Haiku):

• Flowchart
• ER diagram
• UML Class
• Sequence diagram
• Gantt chart
• Mind map
• Bar chart
• Pie chart

Export as PNG or SVG. Edit any diagram by clicking Apply changes with a plain-English instruction.

12. Stats Analysis (Student & Postgrad Pro)

The Stats tab is a four-step wizard for statistical projects. Supports SPSS, R, Excel, and manual analysis. Use it if your project has hypothesis testing, correlation, regression, ANOVA, or similar quantitative work.

1. Test Recommender (1.5 cr) — paste your research question and hypotheses, pick your measurement scale and tool. The AI recommends the right test (Pearson, Chi-square, ANOVA, etc.), with rationale, assumptions, and step-by-step instructions for your tool.
2. Data Preparation (1.5 cr, optional) — list your variables and the AI generates a data-entry guide tailored to SPSS / R / Excel.
3. Interpret Output (5 cr) — upload a screenshot of your SPSS / R / Excel output. The AI reads it, tells you which hypotheses are supported, and drafts academic prose ready for Chapter 4.
4. Write Analysis Chapter (7 cr) — stitches everything together into a finished Results & Discussion (or Chapter 4) section.

13. Integrity & Defensible Draft

Two tabs handle quality control before submission:

Plagiarism tab — Integrity Check

• Originality scoring (4 credits per section) — compares your section to a similarity index and flags passages that look too close to web sources or other papers.
• AI-pattern detection (Elite / Postgrad Pro) — checks for the linguistic fingerprints of AI-generated content. If a passage scores high, paraphrase or rewrite it before your supervisor uses a detector.

Defensible tab — Defensible Draft

Postgrad Pro only. Stronger guarantees before you submit:

• Originality preflight (0.8 cr per section) — runs before every section export. Block-by-block similarity check.
• Paraphrase coach (0.3 cr per snippet) — for flagged passages, suggests rewrites that preserve meaning while changing structure.
• Citation completeness (0.4 cr per section) — finds claims that should be cited but aren't.
• DOI verification (0.1 cr per reference) — confirms every DOI in your reference list actually resolves.
• Defensible Draft badge — once all checks pass, your project earns a badge you can show supervisors.

14. Defense Simulator

Practise your defense before the real one. The simulator role-plays your project examiner using your actual project content as context.

• Supervisor mode — one AI supervisor, conversational. Good for early practice.
• Panel mode (Postgrad Pro) — three examiners with distinct personalities (the methodologist, the literature-purist, the practical one) firing questions from different angles.
• Difficulty levels — Standard, Hard, Brutal. Pick higher to surface the questions your real panel might ask.
• Each defense turn costs 4 credits. Sessions are limited on Student tier (6 / month) and unlimited on Postgrad Pro.

After each session you get a scored breakdown of every answer — what was strong, what needs work, and the actual question you should have answered.

15. Exporting your project

When the last section is Accepted (or Edited), click Export in the top bar. Export is free of credits.

• DOCX — all plans. Fully formatted with your reference list at the end.
• PDF — Student plan and above.
• LaTeX — Postgrad Pro only. For supervisors who require it or for camera-ready submission.

16. Editing & deleting projects

Projects aren't frozen after creation. You can rename the title, rewrite the abstract, update your supervisor or deadline, or delete a project outright.

From the project list

On /project-lab/app, each project card has a three-dot menu (top right of the card). Click it for two actions:

• Edit details — opens a small dialog where you can change title, abstract, department, supervisor, or deadline. Other fields (institution, project type, academic level) are fixed at creation; only these five are editable.
• Delete project — opens a confirmation. You must type the word delete before the red button activates. This prevents accidental clicks from wiping weeks of work.

From inside the workspace

The top bar also has a pencil icon (edit) and a trashicon (delete) right after the project-title breadcrumb. Same dialogs, same behaviour. Useful when you're already working on a project and want to fix the abstract without leaving the editor.

Why isn't there an undo?

We haven't built a soft-archive flow yet. If you want the safety net, export the project to DOCX (free, no credits) before deleting — you keep a local copy and any references inside it stay readable.

17. How to get maximum results

Project Lab does the heavy lifting only when you give it good ingredients. These habits separate the students who finish in three weeks from those who fight the AI for three months.

Upload before you generate

The single biggest difference in output quality: upload your sources first, then generate. A chapter generated from your actual lecture PDFs, journal articles, and past supervisor notes is dramatically tighter than one generated from a thin abstract alone. Aim for at least 5–10 sources before generating Chapter 2.

Write a sharp abstract

The abstract you put in at project creation is the AI's north star for every section. Vague abstract → vague chapters. Spend 15 minutes writing two clear paragraphs: what is the problem, what are you doing about it, what do you expect to find.

Iterate, don't regenerate

A new section is 7 credits. An improvement is 1.5. After you accept a draft, prefer Improve / Expand over Generate-from-scratch. Same outcome, ~4× cheaper.

Use Improve with specific feedback

"Make it better" is wasted credits. "Cut the second paragraph in half, integrate the Adeyemi 2021 citation more prominently, and end with a transition to methodology" gets you exactly what you want in one pass.

Tag your source quality

Run Auto-tag (in Sources → Library Health) so the AI knows which sources are T1 peer-reviewed vs T5 blog posts. Citations get weighted: the AI leans on T1 for serious claims and de-emphasises T4–T5.

Run the integrity check before your supervisor does

Catch similarity issues before submission. 4 credits per section is cheap insurance against a supervisor rejection or a Turnitin flag. On Postgrad Pro, Defensible Draft runs preflight automatically.

Practise defense weekly, not the day before

One simulator session a week from when you start Chapter 3 onwards. By the real defense, you'll have heard every variant of every question. Panel mode (Postgrad Pro) is closest to the real thing.

Save your monthly allowance for the heavy chapters

Chapters 2 and 4 (Literature Review and Results) eat the most credits. Chapter 1 is mostly scaffolding you write yourself. Plan generation timing across the month so you don't run out two weeks in.

Defense Sprint, used at the right moment

If you're on Student tier and your defense is within 30 days, the ₦12,000 Defense Sprint unlocks Postgrad Pro features (panel mode, Methodology Advisor, Literature Engine, Defensible Draft) for the period you need them most. Time it for the final 4 weeks before defense.

18. FAQ

What happens when I run out of credits?

You can keep editing and reading — only AI-generation actions are blocked. Top up from ₦2,000 or wait for the next monthly refill at the start of your billing period.

Can I work on more than one project?

One active project at a time on every tier. Archive a finished project to start a new one — archived projects stay readable but don't count against the limit.

Is my project data private?

Yes. Project content is encrypted at rest and never used to train AI models. Only you (and anyone you explicitly share with) can read it.

What if the AI cites something that doesn't exist?

It shouldn't — every citation traces back to a source you uploaded. If you spot one that doesn't resolve, click the citation marker to see the source the AI claimed to be using, and report it via the workspace feedback button. Then run the section through Fact-check (0.5 credits) to flag any other unsupported claims.

Can I switch citation styles after writing?

Yes — open the Refs tab and pick a different style. All inline markers and the reference list reformat instantly. No re-generation required.

What if my supervisor rejects a chapter?

Open the section, paste their feedback into the AI tab's Improve box, and click Improve section. The AI rewrites with the feedback applied. Total cost: 1.5 credits.

Can I cancel anytime?

Yes. Subscriptions cancel at the end of the current billing period — your access stays live until then.