The Study API provides four AI-powered content generation endpoints that turn raw course material into active learning tools. Paste or upload notes, and the AI generates flashcards, key concepts, and practice quiz questions — all in one call. A separate evaluation endpoint semantically scores a student’s answer against the correct one, applying lenient partial-credit logic that rewards genuine understanding even when phrasing differs. An image analysis endpoint extracts all text, formulas, diagrams, and captions from a photo of handwritten notes or a textbook page. Finally, per-note summarization condenses uploaded documents into bullet points and key takeaways. These endpoints power the Study & Learn section of IntelliPlan and are available to any authenticated session.Documentation Index
Fetch the complete documentation index at: https://mintlify.com/UAnirudh/IntelliPlan/llms.txt
Use this file to discover all available pages before exploring further.
The study endpoints are web-layer routes (not under
/api/v1/) and use session-based authentication. They are available to any logged-in IntelliPlan user. Guest users have access to a limited quota (capped generations and fewer questions).POST /study/generate — Generate Flashcards and Quiz
Takes a block of study text and returns a structured study set: a title, 5–8 key concepts with definitions, and up to 20 practice quiz questions across three types — recall, conceptual, and short-answer. Content is capped at 20,000 characters; longer text is silently trimmed to that limit.- Request
- Response
cURL
Request Body
The study material to generate content from. Plain text, markdown, or extracted note text. Maximum 20,000 characters (text beyond this limit is trimmed).
Number of quiz questions to generate. Minimum
1, maximum 20. Defaults to 8. Values outside this range are clamped automatically.Study mode intensity. Accepted values:
"casual", "exam", and other configured modes. "casual" is lighter; "exam" generates harder, more comprehensive questions. Guest users are locked to "casual".Response Fields
A short (5 words max) topic title inferred from the content.
5–8 key concepts extracted from the material. Each entry has a
term (string) and definition (1–2 sentence string).Array of quiz question objects.
The study mode actually used for this generation (may differ from the requested mode if the user was downgraded from the guest limit).
POST /study/evaluate — AI-Evaluate a Quiz Answer
Semantically evaluates a student’s free-text answer against the correct answer. The evaluation is intentionally lenient — it rewards approximate understanding, partial credit for capturing the main idea, and deducts for high-confidence wrong answers. The engine combines LLM semantic scoring with keyword overlap and sequence similarity checks so that a student who genuinely understands the concept is never penalized for imperfect phrasing.- Request
- Response
cURL
Request Body
The quiz question text.
The model correct answer for the question.
The student’s free-text response to evaluate.
The student’s self-reported confidence level. Accepted values:
"high", "medium", "low". Confidence modifies the points multiplier — a correct high-confidence answer earns more; a wrong high-confidence answer earns less. Defaults to "medium".Response Fields
Outcome:
"correct", "partial", or "incorrect". The engine is lenient — if the student captures the main idea, the verdict is at least "partial".Numeric score from 0–100. Correct answers score 70–100; partial 40–69; incorrect 0–39.
Encouraging feedback on what the student got correct.
Precise description of any gaps or misconceptions.
2–3 sentence constructive critique of the answer.
A vivid mnemonic or analogy to help the student remember the concept.
A concise ideal answer the student can use as a reference.
Gamification points awarded for this answer (1–15), factoring in verdict and confidence.
POST /study/analyze-image — Extract Content from Image
Uploads an image of a textbook page, whiteboard, handwritten notes, diagram, or any educational material and returns all extracted text, formulas, tables, and visual element descriptions as clean study-ready text. The AI vision model preserves mathematical notation, describes diagrams, and retains all labels and captions.- Request
- Response
cURL — multipart upload
Request
Send amultipart/form-data POST with the image file in a field named image.
The image file to analyze. Supported types:
image/jpeg, image/png, image/webp, image/gif. Maximum file size: 10 MB.Response Fields
All extracted educational content from the image, formatted as clean readable text. Mathematical formulas are preserved; visual elements are described in square brackets.
Character count of the extracted text, useful for deciding whether to pass the result directly to
/study/generate.POST /notes/{id}/summarize — Summarize Uploaded Notes
Generates an AI summary of a previously uploaded course note file. The summary is structured as 5–8 bullet points plus a short “Key takeaways” section. The generated summary is cached on the note record — subsequent calls to this endpoint will regenerate it if the note’s text content has changed.cURL
The integer ID of the course note to summarize. Note must belong to the authenticated user.
200 OK
The AI-generated summary as a markdown-formatted string with bullet points and a Key takeaways section.
Image-to-Flashcards Workflow
Photo to text
Upload a photo of handwritten notes or a textbook page to
POST /study/analyze-image. Receive clean extracted text.Text to study set
Pass the extracted
text to POST /study/generate as the content field. The AI returns key concepts and quiz questions.