Why 300,000 students chose Fabric over ChatGPT for studying

ChatGPT answers from the internet. Fabric answers from your lecture notes. For students, the difference is the difference between a generic tutor and one who's taken your course.

ChatGPT is the most used AI tool among university students. It's also the most misused. Students paste in their essay questions and get back polished text that may or may not be accurate, may or may not match their course material, and definitely doesn't cite the sources their professor assigned. The AI is smart. It just doesn't know their course.

Fabric's student adoption has grown rapidly because it solves a specific problem that ChatGPT doesn't: grounding AI answers in the student's actual course materials. The distinction sounds subtle until you experience it. "Explain the Krebs cycle" in ChatGPT produces a textbook answer from the internet. "Explain the Krebs cycle using what was covered in this week's lecture" in Fabric produces an answer drawn from the student's recorded lecture, citing the specific timestamp where Professor Chen explained the electron transport chain.

The citation is the difference. The grounding in your materials is the difference. The AI that knows your course is categorically more useful for studying than the AI that knows everything except your course.

What students do with Fabric

Build a searchable library across their entire degree. Every lecture recording is transcribed and searchable. Every PDF (readings, slides, past papers) is indexed. Every note is searchable by meaning. Every annotated reading is findable by concept. In third year, a student can search across everything from first year. The question for the final dissertation draws on the full accumulated library.

Ask the AI tutor questions grounded in their materials. "What did the readings say about institutional bias?" searches the annotated PDFs and produces an answer with citations to specific pages. "Quiz me on the last three weeks of microbiology" generates questions drawn from the lecture transcripts and assigned readings. The tutor knows the course because it has access to the course materials, not because it was trained on generic textbooks.

Prepare for exams with source-grounded revision. The revision session draws from everything: lecture recordings, annotated readings, notes, and study materials. "Summarise the key themes from this module" produces a synthesis across every relevant source in the library, with citations. The student can verify every claim by clicking through to the original source, which builds confidence that the revision is accurate and comprehensive.

Capture ideas and connect concepts. Voice memos captured on the walk home from the seminar are transcribed and searchable. Quick notes typed between lectures are findable alongside the formal reading notes. The connections between modules that students rarely notice, because the notes are in separate folders, surface through semantic search that understands conceptual relationships.

Collaborate in study groups. Shared spaces let study groups pool their materials. Everyone's notes, readings, and recordings become searchable across the group. The group's AI tutor draws from everything the group has collected, which produces richer answers than any individual's library alone.

Why not just ChatGPT?

ChatGPT is a brilliant tool for many things. For studying specifically, it has three limitations that matter.

It doesn't know your course. ChatGPT answers from its training data, which is the internet. Your course materials, your professor's specific framing, your assigned readings, your lecture content: none of this is in ChatGPT's training data. The AI can explain photosynthesis. It can't explain photosynthesis the way your biology professor explained it, using the specific model and terminology from the assigned textbook.

It doesn't cite your sources. ChatGPT produces answers without citations to your course materials. For exam preparation, you need to know which lecture covered a topic, which reading discussed an argument, and which page contains the evidence. Fabric's citations link every claim to the specific source in your library: the lecture timestamp, the PDF page, the note you wrote.

It doesn't accumulate. Every ChatGPT conversation starts fresh (memory helps with personal facts but doesn't maintain a searchable library of academic content). Fabric's library grows with every lecture, every reading, every note. The AI in month six has dramatically more context than the AI in month one. By final year, the accumulated library is the most comprehensive study resource the student has ever had, and the AI can draw on all of it.

The Bloom's two sigma angle

Benjamin Bloom's 1984 research showed that one-on-one tutoring produces a two standard deviation improvement in student performance. The problem was always cost: personal tutoring doesn't scale. A human tutor costs $40-80 per hour and is available a few hours per week.

Fabric provides an AI tutor that's available at any hour, grounded in the student's actual course materials, and costs $10/month. It's not a human tutor. It doesn't provide the mentoring, motivation, and relationship that a great human tutor offers. But it provides the personalised, source-grounded, always-available academic support that the two sigma research identified as transformative, at a cost that makes it accessible to every student rather than only those who can afford private tutoring.

That accessibility is why 300,000 students have chosen it.

Frequently asked questions

Is using Fabric for studying considered cheating? Using Fabric to understand your course materials, find relevant sources, and prepare for exams is studying, not cheating. The AI helps you learn from your materials. It doesn't write your essays or do your homework. Many universities explicitly encourage AI tools for study support while prohibiting AI for submitted work.

Does the AI tutor replace going to lectures? No. The AI tutor works best when you've attended lectures because the recorded lectures feed into your library. The tutor draws on your lecture content. Without the lectures, the library is thinner and the AI has less to draw on. Fabric makes lecture attendance more valuable, not less.

How does Fabric compare to NotebookLM for students? Both ground AI answers in your sources. NotebookLM is bounded (50 sources per notebook, answers only from uploaded content). Fabric is cumulative (unlimited library, answers from everything you've ever added). NotebookLM is better for focused, bounded study sessions. Fabric is better as a persistent knowledge base across your entire degree.

What about Khanmigo for studying? Khanmigo is excellent for K-12 maths and science with its Socratic approach. For university-level studying across all subjects, Fabric's strength is that it works with your specific course materials rather than a pre-built curriculum.

Can I import my existing notes? Yes. Import from Notion, Google Drive, Dropbox, and other tools. Your existing notes become searchable by meaning alongside new content.

Does Fabric work on mobile? Yes. The mobile app handles capture (voice memos, quick notes, screenshots) and search. Capture on your phone, study on your laptop. Everything syncs.

Is $10/month realistic for students? Compared to a single hour of human tutoring ($40-80), a month of Fabric provides thousands of interactions with an AI grounded in your materials. The free tier covers basic features if budget is the primary concern.

What subjects does it work for? Any subject where you have course materials (lectures, readings, notes). STEM, humanities, social sciences, law, medicine, business. The AI works with whatever content you provide.

Related reading: The student second brain, Bloom's two sigma problem, AI that cites your sources, 10 AI study tools replacing tutoring. Related pages: For students, AI tutor, Search, Audio and video transcription.