What it is
mochi.js is a Bun-native, raw-CDP browser automation framework designed to create relationally-coherent fingerprints. It aims to leave no measurable fingerprints by unifying various aspects of browser behavior into a single, consistent stack.
How it works
The framework operates on five pillars to ensure stealth and consistency. The relational consistency engine derives all fingerprint surfaces—including canvas, WebGL, audio, fonts, MediaDevices, and WebGPU—from a single profile and seed pair using a 48-rule Directed Acyclic Graph (DAG). Additionally, it utilizes a Chromium-native fetch mechanism that routes requests through Chromium via CDP, ensuring that JA4/JA3/H2 values are authentic Chrome identifiers.
Behavioral synthesis is handled through biomechanical models, synthesizing actions like humanClick and humanType using Bezier paths, Fitts-law movement times, and lognormal digraph delays. This approach replaces traditional, hand-stitched pipelines with a single, end-to-end library built exclusively with Bun.
TWSC experience
Not yet tested by TWSC.
Related
Sources
The Web Scraping Wiki is compiled from the tests and research published on The Web Scraping Club, the newsletter for web scraping professionals. Subscribe to get the next article in your inbox.