Our platform makes it possible for other service providers to integrate parts of the Examplary experience within their products. To make this possible, we offer the Embed Sessions API, which allows the service provider to get a unique link to a part of the Examplary UI that can be embedded in their UI.
This makes it possible to integrate flows such as exam generation or question generation directly into your own application, while still leveraging the full power of the Examplary platform.
Getting started
To get started testing this flow, first create a Examplary account.
You'll need to get your organisation approved for using embed sessions. Please contact us at hi@examplary.ai with your organisation details or use the chat widget in the app to request access.
Find your API key in the Developer settings page. Then, follow the instructions in the exam generation flow document to create your first embed session.
Or have a look at the live code example.
Supported flows
- Exam generation (
generate-exam) - Edit rubric (
edit-rubric) - Question generation (
generate-question) - Mark answer (
mark-answer) - Edit exam (
edit-exam) - Take exam (student) (
take-exam) - Practice space (student) (
practice-space-student)
Actor field
Every POST /embed-sessions request requires an actor field: the user ID of the Examplary user who will interact with the embedded UI. Three constraints apply:
- Must be a user ID — the value must start with
user_. - Must differ from the API key owner — the actor cannot be the same user as the one whose API key you are using to authenticate the request. If they match, the API returns a
400error with the message: "Actor ID cannot be the same as the requesting user. Create a new user in the workspace and use its ID." - Must exist in the workspace — the actor must already be a member of your workspace. If the user ID doesn't match any user in your workspace, the API returns a
400error with the message: "Actor with the given ID does not exist as a user within your workspace."
The typical pattern is to call POST /users once per end-user and store the returned user_… ID. Use that stored ID as the actor for all future embed sessions for that user.
Example flow
Example flow for exam generation using embed sessions: