Skip to content

Make it with AI.
Share it as a canvas.

Give your report, prototype, or app a stable link your team can open through your organization’s sign-in. Each canvas is a live web page or app, with named viewers and editors and a history of published versions.

Self-hosted and MIT licensed. Connect your AI agent.

Let your agent publish the canvas

Bring HTML or files from an AI chat, cloud workspace, or laptop. Your agent can create and update canvases over MCP with the same permissions you have in the dashboard.

Keep the link as your canvas changes

Work on a draft while colleagues use the published canvas. Preview and publish updates at the same URL, or make an earlier version current again.

Use your organization’s sign-in

Give people and teams viewer or editor access to each canvas. Open it to your whole organization, or publish a public link when your admin allows it.

Choose who can use
and edit your canvas.

Start with the people and teams who need your canvas. Add them as viewers or editors, then decide whether to open it to your whole organization or anyone with the link.

The people and teams you add keep their access when you change General access. Invite colleagues before their first visit; they use your organization’s sign-in, with no separate password to manage.

How canvas sharing works
People & teams

Add colleagues and teams as viewers or editors. Their access stays in place whichever option you choose below.

Restricted

Nobody else. With no one added, that's just you.

Whole org

Anyone signed in to your organization.

Public link admin-controlled

Anyone with the URL, subject to any password or expiry you set. Available when your admin allows it.

Give your canvas
the backend it needs.

Six built-in capabilities let your canvas save data, accept uploads, call AI and work live with other people. Use them through the browser SDK after enabling access; authoring is a separate opt-in.

Data storage

Save shared data and each viewer’s preferences. Group authored records into collections with their own permissions.

Files

Let people upload files to a canvas. Set who can read or change them; attachments follow their parent record’s access.

AI

Add AI responses using your configured models and quotas. Provider keys stay on the server.

Identity

Use each viewer’s signed-in identity, canvas role and permissions. The server resolves who they are.

Realtime

Add live messages and presence. Choose who can receive, publish and appear in each channel.

Connections

Call approved third-party APIs through admin-granted Connections. Credentials stay on the server.

Authoring

Let signed-in viewers create and manage canvases as themselves. Off by default, instance-wide and per canvas.

Make room for canvases
across your organization.

Use the sign-in your organization already has. Keep ownership clear, review access, and hand over canvases when people change roles.

Teams & invites

Share with a team as viewers or editors. Invite colleagues by email before their first sign-in. Your admin decides whether people outside your organization can be invited.

Editors & ownership

Editors help manage, edit, and publish a canvas. Deletion, ownership transfer, and enabling guest AI remain owner-only. Owners can hand work to an editor; admins can reassign it when someone leaves.

Org sign-in (SSO)

Use your OIDC provider or identity-aware proxy, with your allowed domains and people. Canvas Drop does not manage user passwords.

Admin console

Filter by owner, access, or status and save useful views. Inspect ownership and permissions without gaining access to private canvas content.

Your brand, your look

Choose from four design skins for your dashboard, editor, and public pages. Change the whole instance from the admin console without a restart.

A clear handover

Preview a departing person's canvases and access. Reassign ownership, block the account, revoke sessions and agent tokens, and see exactly what still needs follow-up.

Investigate and act

Check why someone has access, review administrative activity, and preview bulk actions. Restore deleted work during retention or purge it afterward; backups follow their own policy.

Connections you can operate

See recent failures, latency, and affected canvases. Test a connection and rotate its credentials while their values stay on the server.

Run Canvas Drop
on your infrastructure.

The source is yours under the MIT license. Run the Docker image on a VPS or in your cloud, with your choice of database and file storage.

Restricted by default

New canvases start Restricted. Only the owner and the people or teams they add can open them until access is widened.

Usage stays in your instance

Canvas Drop does not phone home or load third-party tracking. Product usage data stays in your instance, and fonts are served locally.

Secrets stay server-side

Built-in AI and Connections keep provider credentials on the server. The backend resolves each visitor’s identity and permissions for every request.

Backend off by default

You choose which backend capabilities a canvas may use. Named collections, file groups, channels, and Connections each have their own permission controls.

Your infrastructure

Choose SQLite or Postgres and local disk or S3 for your canvas files and data. Configured AI providers and Connections use the external services you choose.

Make your next creation
a canvas people use.

Bring your HTML or files, or connect your AI agent. Publish a canvas and share its link with your team.