Use AI Studio in a canvas
Create AI output from a canvas you can edit, use Object Chat, review results, manage credits, and recover failed or pending AI work.
AI Studio belongs inside the canvas workflow
Use AI Studio in a canvas when the output should become part of the work people are already reviewing. You can generate a draft, table, image, audio result, webpage-style result, or other supported output from a prompt, selected canvas objects, attached files, pasted links, or voice input where available.
The canvas matters because AI output is not just text in a chat box. ALLO can place a pending result on the canvas, let you preview and edit the finished output, and let teammates review it with comments, Team chat, or sharing. Use AI Studio when you want AI help that stays close to the source material and the team’s review space.
For the broader AI Studio article, read Use AI Studio. For selected item questions, read Ask about a canvas item.
Availability and permissions
| Item | Details |
|---|---|
| Available on | Workspaces and plans with AI Studio enabled. |
| Available for | Web app and desktop app in supported canvas editor contexts. |
| Who can use it | Workspace members with AI access, available workspace AI credits, and permission to edit the canvas. |
| Where it appears | In an editable, non-embedded canvas. |
| Not available to | Guests, external collaborators, view-only access, comment-only access, embedded canvases, or workspaces without AI Studio. |
| Credits | AI work starts only after the workspace credit check passes. Admins manage credits in Billing. |
Where AI Studio appears inside Canvas
Open a canvas where you have edit access. AI Studio is the canvas AI panel for creating new output from the current canvas, current viewport, selected objects, files, links, and prompts.
If you are a guest or external collaborator, AI Studio is not available even when the shared canvas lets you view, comment, or edit. Ask a workspace member to generate the output. If you are a workspace member and AI Studio is missing, ask the canvas owner for edit access or ask a workspace admin to confirm AI Studio and credit availability.
Generation modes
AI Studio can start from different kinds of input. Pick the mode that matches your source material.
| Mode | Use it when | Example |
|---|---|---|
| Prompt only | You know what you want and do not need canvas context. | Create a launch checklist for a customer webinar. |
| Selected canvas objects | Existing notes, files, images, tables, or generated objects should shape the output. | Select research notes and ask for a summary. |
| Attached files | Source material is not already on the canvas or needs to be attached to the draft. | Attach a PDF brief and ask for a client-ready outline. |
| Pasted links | A safe URL should be used as source context where supported. | Paste a public reference link and ask for a comparison. |
| Voice input | You want to dictate the prompt. | Record a rough request, then review the transcribed prompt before submitting. |
| Object Chat handoff | You are asking about a specific item and may want follow-up output. | Ask Object Chat to critique a generated table, then create a revised version. |
AI Studio chooses different handling based on your intent and context. It can place uploaded source material on the canvas, open or continue a chat about a source item, or generate a new output object. Review the visible action before submitting.
Attachments and source context
Attachments can make AI Studio much more useful, but only when the source material is relevant and successfully uploaded. Where supported, AI Studio can work with common images, PDFs, document files, plain text, Markdown, CSV, JSON, SRT, and spreadsheet files. The available file types can vary, so use the upload UI as the source of truth.
Keep attachment sets small and purposeful. If you attach too many files or a file is still uploading, failed, unsupported, password-protected, corrupt, or too large, the request may fail or produce incomplete output. Wait for uploads to finish before submitting. Remove failed files and retry.
If the source is already on the canvas, selecting the object can be clearer than uploading another copy. For object-specific questions, use Ask about a canvas item.
On freeform whiteboards, smart arrows can preserve source-to-result relationships. Attach the arrow from the source object to the result object when that relationship should help AI Studio or Object Chat understand what came before.
Output types
AI Studio can create several kinds of canvas output where enabled.
| Output | Use it for | Canvas behavior |
|---|---|---|
| Document or text-style content | Briefs, notes, plans, summaries, proposals, explanations. | Appears as a canvas object you can review and edit where supported. |
| Table or structured content | Matrices, comparisons, extracted data, planning lists, scoring sheets. | Appears as a structured result or table-like object. |
| Image | Concept visuals, draft graphics, moodboards, variations, illustrative assets. | Appears as an image result after generation. |
| Audio | Audio drafts or playable media where supported. | Appears as an audio result or object after processing. |
| Webpage-style output | Page concepts, interactive drafts, landing-page style results, structured previews. | Can expose preview and share behavior before reuse or handoff. |
| Chat answer | Explanation, critique, rewrite, summary, or follow-up from Object Chat. | Can be reused manually, turned into canvas content, or used to guide a new generation. |
Use the output type that matches the final artifact. Asking for a table inside a document prompt can work, but choosing a table output where available makes review easier.
Pending outputs and placeholders
When you submit a generation request, ALLO can place a pending output placeholder on the canvas near the current working area. The placeholder is the visible sign that AI work is in progress.
During generation, some actions are disabled. You can select, move, resize, align, comment on, or delete a placeholder, but preview, download, replace, edit, and AI follow-up actions stay unavailable until processing finishes.
Deleting a pending placeholder can cancel or abandon the generation. If you want the result, do not delete the placeholder just because it is taking time. Refresh first if the state looks stale.
If you leave the canvas after ALLO accepts the request, long-running generation can continue server-side. Return to the canvas to review the result. If the canvas itself is deleted while generation is running, the pending work may be cancelled or unrecoverable.
Preview and reuse output
After generation finishes, preview the result before sharing it. Check source accuracy, formatting, tone, sensitive content, and whether the result belongs in the shared canvas.
Useful reuse options include:
| Reuse path | Use it when |
|---|---|
| Edit the result | The output is close and only needs human judgment. |
| Move it near source material | Reviewers need to understand what source the output came from. |
| Add comments | Teammates should review or approve specific parts. |
| Ask Object Chat | You want critique, summary, rewrite suggestions, or explanation about the result. |
| Generate a new version | The source or format was wrong enough that editing is slower. |
| Share the canvas | The result is ready for teammates or external reviewers with the right permission. |
| Delete the result | It is misleading, wrong, or no longer useful. |
Read Review AI output and share web pages before using AI output in client-facing or decision-heavy work.
Use Object Chat inside Canvas
Object Chat is for questions about a specific canvas item. Select a supported object and use the AI action available for that item. You can ask AI to summarize a document, explain a table, critique a generated result, rewrite a note, extract action items, or identify what context is missing.
Object Chat is different from AI Studio generation. Object Chat helps you understand or transform an item through conversation. AI Studio creates new output on the canvas. Both can be useful in the same workflow: ask Object Chat what is wrong with a draft, then use AI Studio to create a revised version.
Object Chat is also different from Team chat. Team chat is for people. Object Chat is for AI assistance.
Credits and billing behavior
AI Studio, Object Chat, voice transcription, and other AI actions can spend workspace AI credits. The balance belongs to the workspace. The person who starts the action can be recorded for history, but the credits are shared workspace credits.
Before provider work starts, ALLO checks whether the workspace has enough AI credits. If credits are insufficient, admins are sent to Billing or a credit purchase flow. Non-admin members, guests, and external collaborators see a message telling them to ask a workspace admin.
Admins can buy one-time credit packs and configure Auto-reload where Billing allows it. Auto-reload can automatically buy credits when the balance drops below a threshold and stays within the monthly cap. For details, read Understand AI credits and Buy and manage AI credits.
Guest and external collaborator restrictions
A guest or external collaborator may have permission to view, comment, or edit a shared canvas but still not be allowed to run AI Studio. That is expected when the workspace limits AI usage to members, editors, or admins, or when spending workspace AI credits is not allowed for outside users.
If an external person needs AI output, a workspace member can generate it, review it, and share the final canvas content with the external person. Do not broaden workspace membership just to expose AI controls unless the person should become part of the workspace.
Common mistakes
Do not open a read-only shared link and expect AI Studio to appear. Ask for edit access or have an editor generate the output.
Do not attach files and submit before uploads finish. Failed or pending files may not be included.
Do not leave misleading AI drafts in a shared review area. Move, label, edit, or delete them.
Do not regenerate repeatedly when manual editing would be faster. Repeated AI work can spend more credits.
Do not use AI output as final support, legal, billing, or customer-facing language without human review.
Do not use Team chat, Object Chat, and AI Studio interchangeably. They solve different problems.
Failure recovery
If AI Studio is missing, check workspace membership and edit access first. AI Studio is not available to guests, external collaborators, embedded canvases, or view/comment-only access.
If generation fails before submission, check your prompt, selected objects, files, and credit state. Your draft may still be available locally if the request was not accepted.
If a file upload fails, remove it and retry. Convert unsupported or large files to simpler formats if needed.
If output is stuck pending, refresh the canvas and check whether the placeholder is still present. Do not submit the same request repeatedly until you know whether the first request completed.
If the result is wrong, improve the prompt, choose a better output type, select the correct source objects, attach the right files, or ask Object Chat what to fix. Then regenerate only if manual editing is not enough.
If credits are blocked, ask a workspace admin to buy credits or configure Auto-reload. If you are an admin, open Billing from the credit prompt or Workspace settings.
If you contact support, include the workspace name, canvas link, output type, prompt summary, attachment types, approximate time, whether a placeholder appeared, and whether the issue was missing AI Studio, failed upload, stuck pending state, wrong output, or insufficient credits.
Related articles
- Add and select canvas elements
- Connect objects with smart arrows
- Use AI Studio
- Ask about a canvas item
- Review AI output and share web pages
- Understand AI credits