Skip to content

Understanding Token Usage

How AI tokens work, what each action costs, and how to make the most of your monthly allowance.

Every time StudyFlow generates a summary, creates exam questions, or processes your course material, it uses AI tokens. Tokens are the unit of measurement for AI language models — roughly 1 token = 0.75 words. Understanding how tokens are consumed helps you plan your usage and pick the right subscription tier.

💬

What are tokens?

Tokens are the smallest units that AI models process. A single word like "exam" is typically 1 token, while longer words like "understanding" may be 2–3 tokens. StudyFlow tracks two types:

Input tokens

The content you send to the AI — your course PDF text, images, and the instructions StudyFlow uses to generate output.

Output tokens

The content the AI generates back — summaries, exam questions, outlines, quiz answers, and study plans.

📊

Token usage by action

Here's what each StudyFlow operation typically costs in tokens. These are estimates — actual usage varies based on content length and complexity.

Action Input Tokens Output Tokens Total (approx.)
Process a digital PDF (per chapter) 2,000–8,000 1,500–4,000 ~5,000–12,000
Process a scanned/image PDF (per chapter) 5,000–15,000 1,500–4,000 ~8,000–19,000
Process uploaded images (per image) 1,000–5,000 1,000–3,000 ~2,000–8,000
Generate chapter outline 500–2,000 500–1,500 ~1,000–3,500
Generate chapter summary 2,000–6,000 2,000–5,000 ~4,000–11,000
Generate exam questions (per chapter) 2,000–5,000 1,500–4,000 ~3,500–9,000
Generate discussion questions (per chapter) 2,000–5,000 1,000–3,000 ~3,000–8,000
Create a study plan 1,000–3,000 1,000–2,500 ~2,000–5,500
Quiz session (5–10 questions) 1,500–3,000 1,000–2,000 ~2,500–5,000
Chat message (single turn) 500–2,000 300–1,500 ~800–3,500
Custom course generation 500–1,500 5,000–15,000 ~5,500–16,500
📄

PDF types and their token impact

Not all PDFs are created equal. The way your document was created significantly affects token consumption:

🗎

Digital PDF

Most efficient

Created digitally (Word, LaTeX, Google Docs). Text is directly extractable.

  • ✓ Lowest token usage
  • ✓ Fastest processing
  • ✓ Most accurate results
📷

Image-based PDF

Moderate usage

PDF containing embedded images with text (e.g., slides exported as images). Requires AI vision to read.

  • ⚠ 2–3x more input tokens
  • ⚠ Slightly slower processing
  • ✓ Good accuracy for printed text
🖨

Scanned PDF

Highest usage

Physically scanned documents or photographs of pages. Each page is processed as an image.

  • ⚠ 3–5x more input tokens
  • ⚠ Slowest processing
  • ⚠ Accuracy depends on scan quality
Tip: Whenever possible, use digitally created PDFs instead of scanned documents. If you only have a scanned version, check if your university provides a digital copy — it will save significant tokens and produce better results.
💰

Token allowances by subscription tier

Free

100K tokens/month

Split into feature buckets: 50K summarization, 20K exam questions, 15K quiz, 5K study plan and 10K other. Covers two small courses when they fit those buckets.

Pro

Own API key no hard limit

Bring your own AI API key. You pay the AI provider directly for usage — no platform cap. Most cost-effective for heavy users.

Premium

1M tokens/month

Enough for 8–15 full courses per month including all summaries, questions and quizzes. No API key needed.

📚

Real-world examples

Here's how token usage adds up for typical student scenarios:

One digital PDF course (10 chapters)

Premium or Pro

Processing: ~60K | Summaries: ~75K | Questions: ~60K | Study plan: ~4K | 5 quiz sessions: ~20K

Total: ~219K tokens

One scanned PDF course (8 chapters)

Premium or Pro

Processing: ~120K | Summaries: ~64K | Questions: ~48K | Study plan: ~4K | 5 quiz sessions: ~20K

Total: ~256K tokens

Quick try — 1 digital PDF (3 chapters)

Free tier (fits bucketed use)

Processing: ~18K | Summaries: ~20K | Questions: ~15K

Total: ~53K tokens

Full exam period — 5 digital courses

Premium (full allowance)

5 × ~200K average per course + quizzes

Total: ~1M tokens

Tips to reduce token usage

Use digital PDFs when possible

Digitally created PDFs use 3–5x fewer tokens than scanned documents. Ask your professor or check your university portal for digital versions.

Split large PDFs

Use the PDF splitter to upload only the chapters you need instead of a full 500-page textbook. This avoids processing irrelevant content.

Don't regenerate unnecessarily

Summaries and questions are saved. You only need to regenerate if you change your education level or output language.

Monitor your usage

Check your token meter in Account → Subscription. The colour-coded bar shows how much of your monthly allowance remains (green = plenty, amber = moderate, red = running low).

Ready to transform the way you study?

Get AI-powered summaries, exam questions and adaptive quizzes that turn dense PDFs into a complete study system in minutes. Built by students, free to start.

Get Started — It's Free