Example output
Follow one candidate's application through a hiring process. Real saved analyses, in the actual app interface. No account needed.
These examples were run through the Interview Intel pipeline once and saved for this demo. What you see here is the real output, rendered with the same templates used by the full app.
In the app, analysis runs asynchronously. When a user submits an email, the request returns immediately while a Celery worker on a dedicated queue processes it in the background. The worker runs the classifier and roughly six LLM calls, and the page polls a status endpoint until the analysis is complete. That usually takes about 10–15 seconds. For this demo, the analysis has already been completed and saved, so the same flow is shown without running the pipeline again. The “Analyzing…” state uses the real template and poller, but it only waits about 5 seconds before showing the stored result.
Nothing you do here runs the analysis pipeline, writes to the database, or triggers a paid API call.
About this endpoint
Worth being specific, since this is a public unauthenticated route on an app that otherwise costs money per request.
- It's read-only. Every demo URL is a GET. It reads one committed JSON file and renders a template. No insert, no update, no queued job, no email.
- It holds no user data. Every email here was written for this page, including the candidate, the company and the role. Nothing here belongs to anyone, and nothing you do here is stored against you.
- The demo routes do not invoke the ML or LLM services or enqueue analysis jobs. Those calls happened once, when these samples were generated.
- The sample name is matched against a fixed list of committed examples, not used to build a file path. Each step's choices come from that same committed data, so a demo URL cannot reach anything that was not shipped with the code.
- It's rate limited per IP, like the rest of the app. Free to serve isn't the same as free to abuse.
Each step takes about 5 seconds.