Built a research hub and upgraded sable vision

Today I pushed two big pieces into Pinkk. First, the new research hub lets users save text notes, screenshots, images, and pull in quick social media snippets from X and Threads. I wired the UI to allow a single click “Save to Hub” that captures the current page and tags it with the URL and a timestamp. It feels like a lightweight notebook that lives inside the platform.

Second, I upgraded sable vision. The old version only flagged screenshots; now it runs an OCR pipeline on every image in the resource tab, extracts key elements, and auto‑generates a concise summary. It can also crop out irrelevant parts—so if you drop a full‑page screenshot, sable will pick out the headline, author name, and main image. This is the core of the upcoming overview analysis report and will be the heavy lifting for any research session.

I spent the afternoon debugging the image cropper on low‑resolution images and tweaking the confidence thresholds for the OCR. The result is a 30‑second summary that still feels accurate enough for a quick read.

The next step is the Chrome extension. It will let users press a button, automatically capture the visible content, and push it straight into the research hub. I’ve finished the first prototype and plan to ship it next week. The goal is to make research from clicks to compiled insight a few minutes instead of hours.

Overall, it’s a solid day of adding real value for our users, but the work is far from done. I’ll keep refining sable’s autonomy and work on making the extension’s UX as frictionless as possible.