Skip to main content
Creating a Courser chatbot takes a few minutes. Once you’ve added your lecture videos, Courser automatically fetches transcripts, chunks the content, and builds a searchable knowledge base your students can query immediately.

​
Before you start

  • You must be signed in to your Courser professor account.
  • You need at least one YouTube URL for a lecture recording.
  • Videos must have English captions enabled on YouTube. Auto-generated captions work.
YouTube Shorts, playlists, and private videos are not supported. Use standard watch URLs only.

​
Steps

1

Sign in

Go to chatcourser.com and sign in with Google or your email and password.Only professors need an account. Students access the chatbot through a public link — no sign-in required.
2

Open the create form

From your dashboard, click Create Chatbot. A form will appear with two fields: a name and a list of YouTube URLs.
3

Name your chatbot

Enter a descriptive name for your course chatbot — for example, BIOL 201 Assistant or Intro to Algorithms Bot. Students will see this name in the chat interface.You can rename the chatbot later from the course settings page.
4

Add your YouTube lecture URLs

Paste one or more YouTube video URLs, one per line. Courser accepts standard YouTube watch URLs:
https://www.youtube.com/watch?v=fDAPJ7rvcUw
https://www.youtube.com/watch?v=NJSO52hGZGs
Shortened URLs (youtu.be/...), embed URLs, and playlist URLs are not accepted. Courser normalizes all links to the https://www.youtube.com/watch?v=VIDEO_ID format internally.
You can add multiple videos in one go. Each video’s captions are fetched and processed independently.
5

Create the chatbot

Click Create. Courser will:
  1. Fetch captions — Downloads English captions for each YouTube video using the youtube-captions-scraper library.
  2. Chunk the transcript — Splits each transcript into segments of approximately 1,000 characters to ensure precise retrieval.
  3. Generate embeddings — Each chunk is embedded using OpenAI’s text-embedding-ada-002 model.
  4. Store in Pinecone — Embeddings are indexed in Pinecone so that semantic search works at query time.
  5. Create an OpenAI assistant — A dedicated OpenAI assistant is configured with your course context.
Processing takes approximately 30–90 seconds per video depending on length. Do not close the browser tab while processing is in progress.
6

Verify and test

Once processing finishes, you’ll be redirected to your chatbot’s admin page. Click Open Chat to open the public interface.Ask a question based on something covered in your lecture. A well-formed answer with timestamped citations means the chatbot is working correctly.

​
What happens if captions are missing?

If a YouTube video has no English captions — manual or auto-generated — Courser cannot extract its transcript and the video will be skipped. To check whether a video has captions:
  1. Open the video on YouTube.
  2. Click Settings (gear icon) in the player controls.
  3. Select Subtitles/CC.
If no subtitles appear, enable auto-captions in YouTube Studio before adding the URL to Courser.

​
What’s next?

Customize your chatbot

Change colors, background image, placeholder text, and AI behavior

Embed your chatbot

Add the chatbot to Canvas, Blackboard, or any website

Build docs developers (and LLMs) love

Get started for freeTalk to us