Group photoCreate or upload one photo where all speakers are visible together.
→Choose your starting point
How do you want to create your conversation?
Choose the option that matches how you want to bring the people into the conversation.
Everyone stays together in the group photo.That group photo remains the visual source throughout the conversation.
Premium · Creative Director
Turn one group photo into a real conversation.
Create or upload one group photo, describe the topic, then confirm who is who directly on the photo. Creative Director prepares the dialogue and voices; desifaces targets the active speaker turn by turn so the right person receives the right audio and lip-sync.
One group photoIdentify each personVoice for each personSpeaker-by-speaker lip-sync
Topic & dialogueCreative Director turns your topic into speaker-attributed conversation.
→VoicesAssign and approve a consistent voice for every speaker.
→Speaker lip-syncOnly the active person is targeted for each dialogue turn.
→Final conversationJoin the turns into one continuous shared-scene video.
Workflow rule
Group Photo → Voices → Speaker Sync → Final
The group photo stays intact. Confirm each speaker on that photo before video pricing; every dialogue line remains independently reviewable and synchronized to the correct person.
Premium · shared-scene conversation
Director V3Start with the conversation you want these people to have.
Describe the people, topic and desired outcome. Creative Director proposes speaker-attributed dialogue first; the one group photo you create or upload later remains the visual source, so you do not have to build each person separately.
Describe the speakers and their conversation here. Do not paste the group-photo image prompt—the photo is created or uploaded after you approve the conversation plan.
1 continuous group-photo scene
or
Describe the topic and people; Director creates speaker-attributed dialogue for review.
Create or upload one photo where everyone is already together. desifaces does not turn them into separate person images.
Select each participant and click their face so desifaces knows exactly who speaks each line.
Each dialogue turn uses that speaker's voice and mapped position before the synchronized conversation is assembled.