Testing Muse, Grok, Instinct and Manus: How Detectable Are Browser Agents?
LeonixAI · reddit · 2026-10-28
A team tested four AI browser agents — Muse, Grok, Instinct and Manus — against a login demo site to see how much automation footprint they expose:
- Manus: easiest to identify; user agent names Manus and the environment shows virtualization.
- Grok: hardest; a consistent Linux browser/device profile indistinguishable from a human via traditional signals.
- Instinct: reports an NVIDIA GPU but an odd CPU-core/screen-resolution combo.
- Muse: virtualized environment and fixed IP ranges.
No single signal proves automation, which is why the team built a platform combining device, network and behavioral telemetry to classify AI-driven sessions. They invite agent builders (Playwright-based, Qwen, etc.) to test, and pose two questions: how much should agents mimic humans, and should websites be able to detect them at all?
More from coding & agent
- Generating K-Pop music videos with a Claude Music Video skill, no MCP needed — AIandDesign · 2026-10-29
- Pydantic's code mode cuts agent tokens 30% and speeds long tasks 51% across 8 models — samuelcolvin · 2026-10-29
- Decagon details a self-improving autoresearch loop for inference stack optimization — kimberlywtan · 2026-10-29
- Dev uses Claude-built library to ship Obsidian plugin that plans music practice with spaced review — dSebastien · 2026-10-29
- Sitzmann open-sources texwerk, a self-hosted collaborative LaTeX editor with native agent support — vincesitzmann · 2026-10-29
- Professor runs DeckWerk's headless server mode for real-time lab slide collaboration — vincesitzmann · 2026-10-29