Brilliant.org Courses Scraper

Scrape Brilliant.org's interactive learning catalog — math, computer science, data analysis, AI, and science courses. Get full course descriptions, chapter/lesson breakdowns, lesson and exercise counts, prerequisites, and related "next step" courses. No auth, no proxy, no cookies required.

What this actor does

• Two output modes: courses (one record per course, chapters/lessons nested inside) and lessons (flattened, one record per lesson)
• Full catalog discovery: automatically pulls the complete list of ~65 public courses from Brilliant's own sitemap — no manual slug list needed
• Filters: keyword, free-vs-premium access level, lesson-count range, category (learning path / topic taxonomy)
• Rich course metadata: prerequisites, next-step course recommendations, concepts covered, lesson/exercise counts
• Empty fields are omitted

Output per course (mode=courses)

• id, slug, title, description, introText, imageUrl, versionName
• learningPath — Brilliant's topic taxonomy / category slug (e.g. computer-science, foundational-math)
• lessonCount, exerciseCount — course size stats
• premiumOnly, desktopOnly — access flags
• conceptsInclude[] — topic tags covered by the course
• prereqAndNextStepNotes — free-text guidance on prerequisites
• prerequisites[], nextSteps[] — related courses ({title, url})
• collaborators[] — partner organizations credited on the course ({name, link, image}), e.g. Microsoft on "Quantum Computing". Empty on most courses, omitted when absent.
• chapters[] — ordered chapter list, each with slug, name, descriptionBullet1/2, accessStatus (AVAILABLE/LOCKED), locked (bool), lessonCount, and nested lessons[] (slug, title, url, premiumLocked)
• chapterCount
• courseUrl / sourceUrl — canonical course page
• recordType: "course", scrapedAt

Output per lesson (mode=lessons)

• courseSlug, courseTitle, coursePremiumOnly, learningPath
• chapterSlug, chapterName, chapterLocked
• lessonSlug, lessonTitle, premiumLocked
• lessonUrl / sourceUrl
• recordType: "lesson", scrapedAt

Input

Field	Type	Default	Description
mode	string	courses	courses or lessons
courseSlugs	array	–	Exact course slugs to fetch; empty = entire public catalog
keyword	string	–	Substring match on title/description/intro text
accessFilter	select	–	free / premium / any
learningPath	select	–	Category / topic taxonomy filter (13 known values, e.g. computer-science, python, data-analysis)
minLessonCount / maxLessonCount	int	–	Course size range
maxItems	int	65	Hard cap on emitted records

Example: entire free catalog, courses view

{
  "mode": "courses",
  "accessFilter": "free",
  "maxItems": 65
}

Example: lesson-level breakdown of specific courses

{
  "mode": "lessons",
  "courseSlugs": ["how-llms-work", "pre-algebra"],
  "maxItems": 200
}

Example: AI/ML-related courses by keyword

{
  "mode": "courses",
  "keyword": "neural",
  "maxItems": 20
}

Use cases

• Curriculum research — build a searchable index of Brilliant's course catalog by topic and difficulty proxy (lesson count)
• EdTech competitive analysis — track catalog breadth, free-vs-premium mix, and course sizing over time
• Learning-path design — use prerequisites/nextSteps to map recommended course sequences
• Content audits — flatten to lesson level to count total lessons/exercises across the whole platform

Limitations

• Brilliant.org's public catalog totals roughly 65 courses; maxItems above that in courses mode simply returns everything available.
• Individual lessonUrl links redirect anonymous visitors back to the parent course overview page (Brilliant requires a free account to play through a specific lesson) — the URL itself is valid and resolves successfully, it just lands on the course page rather than the lesson for logged-out visitors.
• Lesson content (the interactive problems themselves) requires a logged-in session and is out of scope — this actor scrapes catalog/syllabus metadata only (titles, structure, counts), not paywalled lesson content.
• learningPath is Brilliant's own internal topic-taxonomy slug (13 known values, e.g. foundational-math, computer-science, python); it's not a human-facing category label shown in Brilliant's UI, but it groups courses consistently and can be used as a category filter/facet.
• Brilliant doesn't gate whole courses behind Premium — every course currently exposes its first 2 chapters free and locks the rest (chapters[].locked/accessStatus), so premiumOnly is false for every course in the catalog today and accessFilter: "premium" currently returns 0 records. The filter is kept because premiumOnly is a real field Brilliant's API exposes (and may be used for future premium-only courses); use chapters[].locked per-chapter if you need today's actual free/paywall boundary.

FAQ

Why is premiumOnly sometimes missing from a record? It isn't — Brilliant marks every course with a boolean access flag, so this field is present on virtually every successfully scraped course. If you don't see it, the course page failed to embed course data (rare) and the record was skipped.

What's the difference between courses and lessons mode? courses mode gives you one row per course with the full chapter/lesson tree nested inside (best for a catalog overview). lessons mode flattens that tree so every lesson is its own row with course/chapter context columns (best for lesson-level counting or spreadsheet analysis).

Is proxy or login required? No. Every course's syllabus/metadata is rendered server-side on the public course page and readable without an account.

How fresh is the data? Real-time — each run reads Brilliant's own sitemap and course pages directly, and picks up new/updated courses as soon as Brilliant publishes them.