Pictures, voice & quotes
Attach pictures, dictate a message, and ask about a passage you selected.
Pictures
Add pictures to a message from + → Attach files, by dragging them onto the composer, or by pasting a screenshot. They show as previews above the field; remove one with its ✕ before sending. You can send several at once, and keep talking about them over the following messages.
Large pictures are scaled down before they are sent (to about 2048 px on the long side), so a photo straight from a camera does not slow the request. The model needs to support images — most current models do.
Save or copy a picture from a chat with its download button, or right-click it for Save Image As… and Copy Image.
Ask about a passage
Select text in an answer — or in one of your own messages — and an Ask Actovia button appears beside it. Click it and the passage is quoted above the composer; type your question and send. The message shows the quote and your question together. Remove the quote with Remove quote.
Voice
- Dictate — the microphone in the composer records you and turns it into text in the field. The first time, the system asks for microphone access.
- Read aloud — the speaker on an answer's action bar.
Both use OpenAI's speech models with your OpenAI key. Settings → Voice sets
the Speech to Text Model, Text to Speech Model, Text to Speech Voice,
Output Format, Speed and, for gpt-4o-mini-tts, Voice Instructions
(for example "Speak in a warm, friendly tone").
Large files through S3
By default pictures travel inside the request. If you work with many or large files, Settings → AWS S3 lets Actovia upload them to a bucket of yours and send a link instead. The page lists what the bucket needs (public read, CORS, the IAM permissions) before you fill in Region, Bucket, Access Key ID and Secret Access Key.