May 9, 2026
No crumbs, all chaos
Show HN: Mochi.js: bun-native high-fidelity browser automation library
A new bot-hiding browser tool drops, and the comments instantly turn into a roast
TLDR: mochi.js says it can automate a browser while dodging bot detectors by copying real human-like behavior and browser details. The community response was a mix of intrigue and open mockery, with critics questioning the readable docs, the wisdom of revealing the method, and whether the whole pitch actually makes sense.
A new project called mochi.js arrived promising something very spicy: a way to control a browser while looking less like a bot and more like a real person. In plain English, the creator says it can fake all the little signals websites use to decide whether you’re human, from how your browser looks to how your mouse moves. It’s being pitched as an all-in-one replacement for the usual patchwork of tools people use to sneak past bot checks.
But the real fireworks were in the replies, where readers split into two camps: “this is impressive” and “what on earth did I just read?” One commenter absolutely nuked the launch by saying the homepage and docs were “nearly unreadable on mobile,” then twisting the knife with a jab that the pages looked like something an AI wrote without testing. Another asked the question hovering over the whole launch: if the creator publicly documents exactly how the trick works, aren’t they basically handing anti-bot companies a cheat sheet?
Then came the skeptics. One reader shrugged off the fancy claims and said serious sites would just notice the browser is running in a special debug mode anyway. And the funniest hit of the thread was pure confusion: “WTF are you talking about? This is incoherent?” Even nostalgia showed up, with one commenter getting transported back to the early-2000s JavaScript era because the name “Mochi” sounded like an old library. So yes, the tool claims to leave “no crumbs” — but the comment section left plenty.
Key Points
- •The article presents Mochi.js as a Bun-native browser automation framework built directly on the Chrome DevTools Protocol.
- •It says browser fingerprints are generated from a single profile-and-seed pair through a 48-rule DAG to keep attributes consistent across multiple fingerprint surfaces.
- •It states that network requests use Chromium’s own stack via CDP methods, rather than a separate HTTP implementation.
- •It describes human-like automation primitives based on Bezier movement paths, Fitts-law timing, and lognormal typing delays.
- •The project is positioned as a Bun-only all-in-one replacement for a multi-tool stealth automation pipeline including Patchright, fingerprint-injector, Turnstile clicker, and curl-impersonate.