open source · MIT license

panelcall

Rehearse interviews out loud with an AI interviewer. Then read a debrief that quotes your words.

Watch the demo Source code

01

Three formats

15 or 30min

Interview

Paste the job ad and your CV: the questions follow the role.

35min

Case

Slides, tables, charts and code on screen, with errors planted in them.

30min

Open case

One vague problem, on a whiteboard the interviewer can read.

In English, French, Spanish, German, Portuguese or Italian.

02

A debrief that quotes you

“Tell me about a project that went wrong.”

“Honestly, we missed the deadline, but it was mostly the client changing their mind.”

What the interviewer probably concluded: you shift the blame.

A stronger answer, from your own facts“I committed to a date before the scope was fixed. We shipped three weeks late. Now I get the scope signed off before I commit.”

Every quote is checked word for word.
A figure or a name you never gave becomes a [blank].
The debrief judges your answers, never you.

03

It all stays with you

your key: stays on your computer
your voice: goes from your browser to OpenAI, not through panelcall
your interviews: are saved on your disk
telemetry: none, and no account

≈ $0.75

a 15-minute interview, in US dollars, paid straight to OpenAI with your own key

≈ $0.17

the debrief

04

Run it on your computer

uvx --from git+https://github.com/Taelyos/panelcall panelcall

You need uv, Python 3.11 or later and an OpenAI API key with paid credit: GPT-Live is not on the free tier.