How to Share Annotated Images with Your Team

Annotation work rarely happens in isolation. You draw zones, a colleague reviews them, a client approves them, or a remote teammate needs to pick up where you left off. Getting annotations from one person to another is more complicated than it should be.

This article covers the common approaches and when each is appropriate.

The problem with file attachments

The traditional workflow: annotate, export a JSON file, attach to an email or upload to Slack, send to colleague, colleague opens the JSON in their own tool, finds the zones don’t render correctly because their tool uses a different format, emails back asking for a PNG.

The issues compound quickly:

• Format incompatibility between tools
• No single source of truth — multiple exported files circulating with different versions
• No way to tell which version is current
• The reviewer can’t edit and return — they can only comment and ask you to change
• Changes require a new export and a new attachment

Option 1: Shared drive folder (JSON + PNG)

The simplest step up from email. Export both the native JSON and a flat PNG from RegionKit. Put both files in a shared Google Drive or Dropbox folder.

Works well when:

• The team is small and synchronous
• Annotations don’t change frequently
• Reviewers only need to see the annotations, not edit them

Breaks down when:

• Multiple people are editing simultaneously
• You need to track what changed between versions
• The reviewer is a client or external stakeholder without access to your drive

Option 2: JSON in version control (Git)

For engineering teams already using Git, committing zone configuration JSON to the repository has real advantages:

• Full history of every change with commit messages explaining why
• Branches for testing alternative zone configurations
• Code review for annotation changes alongside code review for the inference pipeline

Limitation: The JSON is not human-readable for non-technical stakeholders. A reviewer who needs to verify that zone boundaries match physical spaces can’t meaningfully read a file of coordinates.

Option 3: RegionKit’s publish flow

RegionKit has built-in scene sharing. Click the Publish button in the toolbar to upload your scene to a URL:

https://editor.regionkit.app?scene=abc123xyz

Anyone with this URL can open the scene in their browser — no account, no install. They see the exact annotations you drew, at the correct positions, with layer visibility and styling preserved.

View-only link — the default. Share this with clients, security reviewers, facility managers, or anyone who needs to verify the zone configuration visually.

Edit link — available from the Share dialog. Anyone with this link can open the scene and publish their own changes. The scene tracks a version history so you can see what changed between versions.

Version history

Every publish saves a new versioned snapshot. Open the History panel (clock icon in the toolbar) to:

• See all published versions with timestamps
• Click any version to load it into the editor
• Compare zone configurations over time

This is useful for:

• Reviewing what changed between two configuration reviews
• Rolling back to a previous version if a change introduced a regression
• Demonstrating to an auditor that zone boundaries were reviewed and approved on a specific date

Choosing based on your workflow

Privacy considerations

RegionKit’s publish feature uploads your scene — including the background image and all annotation data — to Cloudflare’s infrastructure. For camera frames containing identifiable individuals or sensitive facility information, consider whether sharing via URL is appropriate for your use case.

If your data must stay on-premises, use the local JSON export and share via your organisation’s own infrastructure. RegionKit’s offline-first design means all annotation work happens in your browser regardless — publishing is always opt-in.

Related: How to Define Regions of Interest for Computer Vision · RegionKit vs CVAT