Working with files

content on every generator accepts a string or a FileContent object. The SDK extracts text from supported files; it does not host uploads or storage.

FileContent

type FileContent = {
  type: "file";
  data: Uint8Array | ArrayBuffer;
  mimeType: string;
  filename?: string;
};

Node Buffer works for data (it is a Uint8Array). Use filename as a fallback when mimeType is generic (for example application/octet-stream).

Two patterns

Pass the file straight into a generator — extraction happens inside the call:

import { createQuiz } from "edu-sdk";
import { readFile } from "node:fs/promises";

const data = await readFile("./lecture.pdf");

const quiz = await createQuiz({
  model: "google/gemini-3.6-flash",
  content: {
    type: "file",
    data,
    mimeType: "application/pdf",
    filename: "lecture.pdf",
  },
  count: 10,
});

Extract first — inspect, cache, or trim text before generating:

import { extractContent, createQuiz } from "edu-sdk";

const { text, pageCount } = await extractContent({
  type: "file",
  data,
  mimeType: "application/pdf",
  filename: "lecture.pdf",
});

const quiz = await createQuiz({
  model: "google/gemini-3.6-flash",
  content: text,
  count: 10,
});

createLearningSet() extracts a file once, then reuses the text for every nested generator.

Supported types (v1)

MIME	Filename fallback
`application/pdf`	`.pdf`
`text/plain`	`.txt`
`text/markdown`	`.md`, `.markdown`

Text-based PDFs only. Scanned or image-only PDFs without OCR are not supported — extraction throws if no text is found.

From an upload

Your app still owns the upload surface:

const file = formData.get("file") as File;
const data = new Uint8Array(await file.arrayBuffer());

await createQuiz({
  model: "google/gemini-3.6-flash",
  content: {
    type: "file",
    data,
    mimeType: file.type || "application/octet-stream",
    filename: file.name,
  },
  count: 10,
});

Errors

Error	When
`UnsupportedContentError`	Unknown MIME and no usable filename extension
`ContentExtractionError`	Empty bytes, empty extracted text, or PDF parse failure
`InvalidInputError`	Malformed `FileContent` or empty string content

All extend EduSDKError. Full details: extractContent and Error handling.