What is Courser?
Courser helps professors create AI assistants for their courses. Each chatbot is trained on the course’s YouTube lecture videos and can answer student questions with direct citations — linking back to the exact timestamp in the lecture where the answer comes from. Professors can embed their chatbot on any course page (Canvas, Blackboard, a course website) using a simple public link. Students interact with the chatbot directly in their browser without needing an account.Quick Start
Create your first AI chatbot in under 10 minutes
Creating a Chatbot
Step-by-step guide for professors
For Students
Learn how to get the most out of Courser
API Reference
Integrate Courser into your own applications
How it works
Professor uploads lecture videos
Paste YouTube links for your lecture recordings. Courser automatically fetches captions and processes the content.
Courser builds the knowledge base
Transcripts are chunked and embedded using OpenAI’s
text-embedding-ada-002 model. Embeddings are stored in Pinecone for fast semantic retrieval.Students ask questions
Students type questions into the chatbot. Courser finds the most relevant lecture segments and generates a cited answer using GPT-3.5-turbo.
Key features
- YouTube lecture ingestion — Automatically fetches captions from any YouTube video
- Semantic search — Vector similarity search via Pinecone surfaces the most relevant content
- Cited answers — Every response includes timestamped YouTube links back to the source
- Embeddable widget — Share a public link that works anywhere — Canvas, Blackboard, or a course website
- Fully customizable — Set a custom chatbot name, color, background image, placeholder text, and AI instructions
- Bring your own key — Use your own OpenAI API key per course, or rely on the platform default
- Conversation export — Download student Q&A history as CSV for analysis
- Email & Google auth — Professors sign in with email/password or Google OAuth
Tech stack
| Component | Technology |
|---|---|
| Backend | Node.js + Express |
| Database | MongoDB (via Mongoose) |
| Vector database | Pinecone |
| AI / LLM | OpenAI GPT-3.5-turbo + LlamaIndex |
| Embeddings | OpenAI text-embedding-ada-002 |
| Auth | Firebase + JWT |
| Frontend | Next.js + Tailwind CSS |
| File storage | Cloudinary |