Media attachments allow you to show images, videos, audio, or documents alongside interview questions to gather visual feedback and test concepts.
Images:
JPEG (.jpeg, .jpg)
PNG (.png)
WebP (.webp)
GIF (.gif)
Videos:
MP4 (.mp4)
Audio:
MP3 (.mp3)
WAV (.wav)
Documents:
PDF (.pdf)
Maximum file size: 200MB per file
An error message appears if a file exceeds the limit
Open question box
Click media upload area
Select file from device
Preview displays after upload (not available for PDFs)
Media shows to participants with question
Preview:
Images display as thumbnails
Videos display with video player
Audio displays with audio player
PDFs show file name only (no preview)
Delete:
Click delete button on uploaded media
Confirmation modal appears if deleting question with media
Media permanently deleted with question
Up to 10 media items per question
Mixed types allowed (images + videos + audio + PDFs in the same question)
Drag-and-drop to reorder
Optional custom labels up to 50 characters (e.g. "Logo A", "Proposed design")
Labels are optional — if you don't set one, the item simply shows without a label
Labels are visible to participants and help you tell options apart in analysis
When a participant uploads an image as part of their answer, Tellet automatically describes what's in it so the description can inform the interview and your analysis. This description is written in the interview language, so it stays consistent with the rest of the conversation.
Images: No hard pixel constraints, but recommend landscape or square at 1200×800 px minimum for clarity on desktop and mobile. PNG or JPEG preferred for static images, GIF for animations, WebP for smaller file sizes.
Videos: MP4 only, keep under 200 MB. Shorter clips (under 2 minutes) work best — participants won't watch long videos mid-interview.
Audio: MP3 or WAV, under 200 MB. Best for audio-based testing (jingles, voice samples).
PDFs: No preview shown to participants (filename only). Best for text-heavy stimuli like ad copy or packaging text. Not ideal when visual impact matters.
2–4 stimuli is the sweet spot for comparison questions. More than that and participants lose focus.
Always use custom labels — makes analysis significantly easier.
Use higher probing levels (2–3) with stimuli questions.
Add a probing guide that explicitly references the stimuli.
Phrase questions to encourage participants to name or describe the specific option rather than positional references ("the first one"), since the AI interviewer doesn't yet track stimulus references.
Right-click and drag-to-save are disabled on images, videos, and the fullscreen image viewer in the participant chat. This deters casual saving, but it isn't foolproof — a determined participant can still take a screenshot or screen recording, so don't show anything truly confidential.
PDFs are not protected: browsers download them natively, and the participant can save / print / share them. Don't upload PDFs of stimuli you don't want a participant to keep.
Audio messages are not visually protected (they use the browser's native audio player).
On mobile, the close button, navigation arrows and page indicator are positioned so they never cover the image itself.
On desktop with very large stimuli, the viewer fits the image to the available viewport without overflow.
Tap-to-zoom and pinch-to-zoom are supported on mobile; arrow keys navigate between stimuli on desktop.
AI interviewer does not yet understand positional references to stimuli (e.g. "I prefer the second one") — not currently supported
Stimulus randomisation (varying order per participant) — not currently supported within a single question (shuffling multiple items attached to the same question). To rotate different stimuli across participants, use separate questions in a randomised Question Group instead — see “Rotating stimuli across participants” below.
No image cropping or resizing built in — prepare media before uploading
Tellet can’t shuffle multiple stimuli attached to one question, but you can still rotate different stimuli across participants (or counterbalance the order they’re seen in) by moving the rotation up a level, to the question itself:
Create one question per stimulus variant — e.g. “What’s your reaction to this design?” with Concept A attached, and a near-identical question with Concept B attached.
Place those questions together in a Question Group.
Enable that group’s randomize toggle. Tellet will then show the questions — and their attached stimuli — in a random order to each participant, so the variant order is rotated across your sample.
Add a Probing Guide that references the stimulus explicitly (e.g. “ask about the design you just saw”), since the AI interviewer doesn’t yet track positional references like “the first one.”
This is the same mechanism used for A/B concept testing — see Discussion Guide Basics for more on how question groups and randomisation work.
Design feedback — show mockups or prototypes and gather reactions
A/B testing — display multiple options and compare preferences
Product demonstration — show the product in use via video for context
Advertisement testing — display ad concepts to test messaging and creative
UI feedback — show interface screenshots and gather usability feedback
Packaging/product testing — display product photos to test physical designs
Questions with media can use any probing level (0-3). For example:
Question: "What's your reaction to this new homepage design?"
Media: homepage-redesign.png
Probing: 3
Probing guide: "First probe should ask about specific elements they mentioned. Second probe should explore whether design achieves intended goal. Third probe should understand if this would change their behaviour."