Harshil Parekh

Hi! Good to see this platform growing as always. I had shared this earlier in the Strivle days, but I’ve since shipped a more complete version with a landing page and public release.

I built a desktop snipping tool that turns screenshots into structured AI input instead of static images. After taking a snip, it runs OCR to extract text and opens an action layer where you can send the context to a local model or external APIs like OpenAI, Claude, or Gemini.

The goal is to close the gap between seeing something on your screen and being able to immediately reason or act on it. I’ve also been experimenting with scrolling capture, screen recording, and step-by-step workflow documentation that can export into structured files with screenshots and actions.

It’s still early, but I’d love feedback from anyone interested in this kind of workflow or building tools around it.

heres the link: snipagent.app

Currently building SnipAgent, a desktop snipping tool that combines screenshots with AI. After taking a snip, it uses OCR to read text (and now images), then lets users run prompts through either a local LLM (currently Qwen 1.5 Coder because thats what my machine can currently handle locally) or their own Claude, Gemini, or OpenAI API keys. Workflow: Snip → Action Menu with preset/custom prompts → AI-generated response.