Cut the slop.
Filter your X timeline with an AI model you choose and pay for directly. Describe what you want to see — or never see again — in plain language.
- Filters in your own words. "Hide engagement bait and crypto threads", or "only keep software engineering". Whatever you'd tell a person.
- Per-author rules. Never filter someone, or always hide them, straight from a post's ⋯ menu.
- Blocked words, matched before the model is asked.
- Thread escape hatch. One click above the reply box shows every comment in a thread, filters off.
- Nothing shown before it's judged. Undecided posts wait behind a placeholder, so nothing slips through and the timeline never jumps.
- Any OpenAI-compatible provider. OpenRouter, OpenAI, Anthropic, or your own endpoint.
There is no Sharp account, no Sharp server, and nothing to sign up for. Your API key lives in your browser and talks to your provider — nobody else sees your timeline, your filters, or what got hidden.
You pay your provider directly. Nothing is collected, measured, or sent anywhere else.
Sharp isn't in the Chrome Web Store yet. Grab the zip from Releases and unzip it, or build it yourself with Bun:
git clone https://github.com/tshmieldev/sharp
cd sharp && bun install && bun run buildThen open chrome://extensions, turn on Developer mode, click Load
unpacked, and pick the folder.
- General › Connection — choose a provider and paste an API key. OpenRouter is the easiest start: one key reaches most models, topped up a few dollars at a time.
- Pick a model. Browse sorts the catalogue by popularity, price, context or measured latency.
- X › Filtering — write what you want filtered, then Apply.
Open X and scroll. ⌥⇧F turns filtering on and off; ⌥⇧T unfilters the thread you're reading.
Found a bug? Open an issue. Building on it? See CONTRIBUTING.md.
Privacy policy · MIT licensed · Copyright © 2026 tshmieldev