Three steps from screen to Claude

Record, work, copy. Aster packages everything Claude needs without the video file.

01

Hit record

Click, drag to select your area. Mic on or off. No cloud account, no setup.

02

Do your work

Browse a spec, flip tabs, click through your app. Aster captures screen text, narration, and screenshots locally at each key moment.

03

Send to Claude

Hit Send to Claude. A context package with the transcript, clicks, on-screen text, and screenshots is copied, and Claude opens. Review it first in the pop-out player with a synced narrative timeline if you like.

Prefer no copy and paste? Connect Aster to Claude with MCP so Claude pulls recordings on its own, or add an Anthropic API key to generate the analysis right inside Aster.

What you actually use it for

Aster detects your session type and shapes the context package accordingly.

Building from a spec

Open your PRD, select the requirements, flip to your live app and click through it. Aster captures the spec text and screenshots side by side.

Result

Gaps identified, tasks ready to action. Formats cleanly for Claude Code.

Customer call summaries

Record a call. Aster transcribes it and tracks what landed, what was questioned, and what was promised.

Result

Next actions, open items, and follow-ups. No manual notes.

Demo walkthroughs

Walk through a feature with narration. Each click saves a local screenshot. You control the pace.

Result

Talking points, demo structure, and key moments labeled. Shareable as a standalone artifact.

Bug triage

Show the broken state, describe what should happen, reproduce the steps.

Result

A clean bug report: steps to reproduce, expected vs actual, severity, inline screenshots.

Code review

Open a PR, narrate your concerns, show the relevant lines. Click to staging to demonstrate the issue.

Result

Specific feedback with line references and fixes. Ready to paste into GitHub.

Design feedback

Open Figma, talk through what works and what doesn't. Aster captures narration and screenshots per section.

Result

Structured feedback with visuals. Ready to post to Slack or send directly.

More context, less waste

Recording tools that upload to the cloud cost money, leak data, and slow you down.

Stays on your machine

All processing is local. Nothing leaves your computer unless you choose to share it.

On-device transcription

Your narration is transcribed locally with Apple Speech, or an on-device Whisper model for sharper accuracy. No upload.

No extra subscriptions

Replaces Loom, Chorus, and Command-Shift-5. If you have Claude, you have what you need.

Context-aware packaging

Aster reads what you're doing and shapes the output for Claude, whether you're building, reviewing, or on a call.

Aster is named after a native wildflower that thrives in its local environment instead of being shipped elsewhere to grow. Local compute works the same way. The more processing that happens on your own device, using models built for your hardware, the less that travels to a data center. That means less exposure and lower cost, and it's the idea Aster is built around. As on-device models improve, so does Aster.