Screenshots that
AI understands.

Annotate what matters, add the missing context, and hand your coding agent a prompt bundle it can actually use.

Local-first desktop app. Open source. No account required.

Imnota workbench with a collection rail, annotated project settings screenshot and inspector with a description and High agent priority.
Markdown handoff.md + .png

# Settings review

## Overall context

Improve clarity and keyboard accessibility in the project settings.

Priority for agent: High

Make Save changes the primary action. Align the buttons with the fields and increase the button contrast.

1 Make the primary action clear.
Actual Imnota interface with example project content. Markdown excerpt shown beside it.

Give the picture a point.

A screenshot shows the symptom.
Your agent still needs the context.

A raw screenshot leaves the important parts implicit. What is wrong? What should change? Which part matters? Imnota helps you make it explicit, right in your local workflow.

Just the screenshot Where should I look?
Example project settings screen. The Save changes button has low contrast and the action row is misaligned.
The screen is visible. The intention is missing.
With Imnota Here. This. Because.
The same example with an indigo rectangle around the action row, an arrow pointing at Save changes and a numbered annotation.
1 Make Save changes the primary action. Align it with the fields.

From seeing it to explaining it.

Four steps. One clear handoff.

  1. 01

    Collect

    Import, paste or drag in PNG, JPEG or WebP screenshots.

    PNGJPEGWebP
  2. 02

    Mark

    Add arrows, lines, shapes, highlights, text, callouts, steps or sensitive-area masks.

  3. 03

    Describe

    Add collection context, screenshot descriptions and agent priority.

    Priority for agentHigh
  4. 04

    Copy

    Export timestamped Markdown and high-resolution PNG prompt bundles.

    .md+.png

Less explaining around the image.
More context inside it.

Close view of the actual Imnota canvas and inspector, showing annotations alongside the screenshot description and agent priority.
Annotations and context, side by side. Example project.
1

Make visual feedback precise

Point directly at the part of a screen that needs attention. Keep editable annotations separate from the original screenshot.

2

Keep project context together

Organize related screenshots into local collections. Add shared context, then archive and restore collections as the work changes.

3

Export a handoff your agent can read

Copy Markdown, image or both. Large collections split automatically into readable prompt bundles.

The receiving editor chooses which clipboard formats it accepts. Separate copy actions are always available.

Local-first, by design.

Your project stays a folder on your machine.

Plain files you can copy, back up and version-control. Keep your screenshots and context together, in a workspace you choose.

  • No account required
  • No cloud storage dependency
  • No built-in AI provider connection
  • No telemetry by default
  • Portable JSON, Markdown and image files
  • Source screenshots are not overwritten
My Project/
  • project.jsonProject metadata
  • collections/
    • 001-collection/
      • screenshots/
      • annotations/
      • exports/
        Markdown + PNG
Your local collection becomes a portable prompt bundle.

Ready when you are.

Install Imnota.

One command to download the latest desktop build. Then make your first screenshot useful.

Browse all releases

macOS / Linux

Terminal
curl -fsSL https://raw.githubusercontent.com/Dytschgo/imnota/main/scripts/install.sh | bash

No Node.js, pnpm, Electron or development dependencies needed.

Download the macOS ZIP

Windows

PowerShell
irm https://raw.githubusercontent.com/Dytschgo/imnota/main/scripts/install.ps1 | iex

No Node.js, pnpm, Electron or development dependencies needed.

Manual build

Clone the source repository and open its folder in your terminal. Use Node.js 20+ and Corepack, then run the packaging command for your operating system.

From the Imnota source folder
corepack enable
corepack pnpm install
corepack pnpm package:win   # Windows
corepack pnpm package:mac   # macOS
corepack pnpm package:linux # Linux

The last three lines are alternatives. Run only the target you need.

A note for macOS. The current release is not Apple-notarised. If macOS blocks it, follow the README opening guidance.

MIT licensed. Yours to inspect.

Built in the open,
designed for real workflows.

An open-source screenshot annotation tool built with Electron, React and TypeScript. Desktop targets for Windows, macOS and Linux. Read the code, build it yourself or contribute a focused improvement.

A few practical answers.

Does Imnota upload my screenshots?

No. Imnota’s core workflow uses local files and has no cloud sync or cloud storage dependency. You decide where to paste or share an exported bundle.

Do I need an AI account to use it?

No. Imnota has no account requirement and no built-in AI provider connection. It prepares context for the assistant or coding agent you already use.

Which operating systems are supported?

Imnota’s development and packaging targets are Windows, macOS and Linux. Check the latest release for available desktop downloads.

What does the export contain?

A fresh, timestamped Markdown and high-resolution PNG set for the current collection. The Markdown includes collection context, screenshot descriptions, agent priorities and text or callout notes. Large collections may split automatically. You can copy Markdown, image or both, depending on what your receiving editor accepts.

Can I build Imnota from source?

Yes. You need Node.js 20+, Corepack and a platform-supported Electron environment. Follow the manual build instructions or the development README.

Is Imnota Apple-notarised?

The current release is not Apple-notarised. The README explains how to open it through System Settings, Privacy & Security, Open Anyway. Read the macOS installation guidance. Do not disable Gatekeeper.