Skip to content

Generate images & videos ​

You can generate images and videos with AI in Liza, preview the result and save it directly to a task. Works in the web and desktop app.

  • Mistral (EU) runs through Liza (managed) – no key of your own needed, counts against your AI credit.
  • OpenAI, Nano Banana (Google) and Black Forest Labs (Flux) each use your own API key (BYOK). Add the keys under Settings → Personal → AI providers: Nano Banana under Gemini, Flux under Black Forest Labs, OpenAI under Codex (desktop app only). The owner can add them centrally for everyone under Settings → Workspace → AI → Providers.

The same applies to videos with Gemini and Black Forest Labs.

How it works ​

There are two ways to open the image dialog:

  1. In a task's AI chat: type /image followed by your description – e.g. /image a duck in comic style – and send it.
  2. In a task's Files box: click "Generate image" at the top of the box.

In the dialog:

  1. Enter a description (what the image should show).
  2. Choose a provider: OpenAI, Nano Banana, Black Forest Labs or Mistral (EU) – and, where there are several, the model.
  3. Pick the format.
  4. Generate – the preview appears after a moment.
  5. Then Save or Discard (or Regenerate for another attempt).

Save & post to chat ​

  • Save stores the image in the task's Files (the Files box on the right). It is stored securely and visible to everyone with access to the task.
  • With the "Post to chat" option, saving also writes a note into the task's chat.

TIP

Saving is only possible when the chat belongs to a specific task (not a whole list). Open the AI chat directly on a task for this.

Providers ​

ProviderModelAccess
OpenAIgpt-image-1your own API key (BYOK)
Nano BananaGoogle Gemini 2.5 Flash Imageyour own API key (BYOK, Google)
Black Forest LabsFluxyour own API key (BYOK)
Mistral (EU)European image AImanaged (Liza credit), processing in Europe

INFO

OpenAI = your own key: OpenAI's image API requires a real API key (sk-…) – the ChatGPT login is not enough for it. Add the API key under AI providers → Codex (desktop app) or have the owner store it as an org key under Workspace → AI → Providers → Codex.

  • Desktop: the locally stored OpenAI key is used.
  • Web/server: a server-side (personal or org-assigned) OpenAI key is used.

Mistral needs no key of your own and runs on your AI credit.

Generate videos ​

Videos are created on a task as well – the finished video lands in its Files. To open the video dialog:

  1. In a task's AI chat: /video followed by your description.
  2. In a task's Files box: "Generate video".

In the dialog you choose the provider, describe the video and set the format (portrait or landscape) and the duration. Optionally you add a start image and an end image – from the images in the task's files or uploaded right in the dialog. The advanced options hold the resolution and, for FLUX, more formats and sound.

ProviderDurationNotes
Google Gemini Omni3–10 secondsResolution 360p to 4K; 360p costs about a third and is good for trying things out. Google usually rejects images showing recognisable real people.
FLUX 3 (Black Forest Labs)5–20 secondsFirst a cheap preview in HD, then finish exactly that take in the chosen resolution (HD to 4K); with or without sound.

A video takes a while. Generation keeps running in the background – you can close the dialog and follow the progress at the top right.

Billing ​

  • Your own key: If a key for Gemini or Black Forest Labs is stored – personally or by the organization – the video runs on it and the provider bills the key holder directly. Your own key always comes first.
  • Via the AI credit: Without a key of your own, Liza books the video from the AI credit. The dialog shows the estimated price beforehand. For Gemini this route offers 360p and 720p, both at the same price.

Coming soon

Videos via the AI credit are currently in beta and will be available to all workspaces shortly. Until then you need a key of your own for videos.