GratisAIfield guide · learning

exam blueprint / sheet 07 · free workflow

Turn a Syllabus Into a Gap-Finding Practice Test — Free

Paste a syllabus or study guide into a free source-grounded notebook, then turn its units into testable prompts and a seven-day weak-area map instead of rereading the whole document.

source path: Google NotebookLM · checked September 12, 2026

free source-grounded notebook pathinstant, inspectable source path
4 units become 8 test promptsnumbers do work in the object
7-day review route, not a score alonesample data is labeled

What this replaces

forces every syllabus unit into a visible coverage map and separates “not recalled” from “not taught,” rather than returning a generic quiz. The result is a small object you can inspect: it carries a count, a context row, and a next action instead of asking you to trust a paragraph.

Tool note. This workflow uses Google NotebookLM. Google’s NotebookLM help pages were checked on September 12, 2026; the official documentation describes source-based chat and a free-user source limit, while availability and limits can vary by account. NotebookLM beat a generic chat window because the source list makes coverage auditable. Runners-up were a generic flashcard generator (too narrow) and a notes summarizer (no recall test). Use sample data first, then decide what information you are comfortable sending to an external AI tool.

The handoff

1. Prepare the source

Use a syllabus, learning objectives, or a chapter outline. Remove student names and private course links before uploading.

2. Ask for a strict map

Prompt the notebook to return JSON with one object per unit: unit, recall_prompt, answer_key, confidence, evidence_phrase, review_day. Ask it to say “not found” when the source does not support an answer.

3. Run the misses here

Paste the same unit list into the blueprint below. The free companion turns those lines into a reusable coverage artifact without calling an API.

Prompt shape for the external tool: return only JSON in this form — {"heading":"…","metric":"…","rows":[{"label":"…","detail":"…","value":"…"}],"next":"…"}. The companion app below uses the same visible result shape locally and makes no AI request.

Try the blueprint board

Load a sample, replace it with your own notes, and inspect the generated object. Sample values are illustrative, not verified facts.

first-load · sample ready

Loading the sample object…

What breaks first

A quiz covers only the loudest unit

Require one row per source heading and check the coverage count before trusting the questions.

The answer sounds plausible but is not in the syllabus

Keep the evidence phrase and source-grounding request. “Not found” is a useful result.

A low score feels like a verdict

Treat it as a routing signal for the next seven days, not a prediction of the exam grade.

Boundary. Learning aid only: it does not know your instructor’s grading policy or guarantee exam performance. Compare answer keys with your course materials or ask your instructor when the source is ambiguous.

Go deeper without losing the free path

The free board is complete on purpose. The companion app adds a tangible file, history, or handoff only after the result is useful: a printable exam packet and saved attempt history; 8 generated rows × about 90 seconds of manual formatting = roughly 12 minutes avoided per packet

open the blueprint companion