It's worth flagging that the homepage and the docs are both nearly unreadable on mobile. If you're going to have cheeky rhetoric like "no crumbs" please take the time to actually test the pages your LLM made for you in a bunch of different ways.
Reads fine for me. I’m getting a bit worn out with the number of ‘if you’re going to use an LLM then take the time to …’ posts. I’ve looked at plenty of HN submissions that don’t render at all on mobile and yes I found it frustrating but didn’t just attribute it to an LLM.
Forget all the fancy fingerprinting - sites that really care about blocking bots will detect that, in the first place, you are running a browser in debug mode. Does this tool prevent that?
> The entire DAG, fingerprint manifest schema, harvesting process, is documented. We even commit our live benchmarks to the public record (mochi on a Linux datacenter IP scored a suspect_score: 8 and bot: not_detected against FingerprintJS Pro v4).
Doesn't this defeats the purpose? Essentially giving away the mechanism to the ones implementing fingerprint so they can learn how to detect you and starting a cat/mouse chase in the open
There's a reason why some kinds of technology stay opaque, not always publishing these things help in the way you think they do
it's not that much of a black box when you can literally see the API's they call(albeit some with more effort than others), but i prefer not to engage in theatre..
"mochi.js is a Bun-native, raw-CDP browser automation framework. Pass a seed and a profile; get a relationally-coherent fingerprint that survives a getParameter(0x9245) probe. JIT-friendly inject payload, Chromium-native fetch (real Chrome JA4 by definition), behavioral synth on top of Bezier+Fitts. Leaves no crumbs. "
It's worth flagging that the homepage and the docs are both nearly unreadable on mobile. If you're going to have cheeky rhetoric like "no crumbs" please take the time to actually test the pages your LLM made for you in a bunch of different ways.
Reads fine for me. I’m getting a bit worn out with the number of ‘if you’re going to use an LLM then take the time to …’ posts. I’ve looked at plenty of HN submissions that don’t render at all on mobile and yes I found it frustrating but didn’t just attribute it to an LLM.
Check the cards that are wrapped, along with the footer
bruh
llm recommends using a desktop to view the site as typically that's also the form factor required to use the tool!
Forget all the fancy fingerprinting - sites that really care about blocking bots will detect that, in the first place, you are running a browser in debug mode. Does this tool prevent that?
Not this but I do have something along those lines if you're interested!
> The entire DAG, fingerprint manifest schema, harvesting process, is documented. We even commit our live benchmarks to the public record (mochi on a Linux datacenter IP scored a suspect_score: 8 and bot: not_detected against FingerprintJS Pro v4).
Doesn't this defeats the purpose? Essentially giving away the mechanism to the ones implementing fingerprint so they can learn how to detect you and starting a cat/mouse chase in the open
There's a reason why some kinds of technology stay opaque, not always publishing these things help in the way you think they do
it's not that much of a black box when you can literally see the API's they call(albeit some with more effort than others), but i prefer not to engage in theatre..
"mochi.js is a Bun-native, raw-CDP browser automation framework. Pass a seed and a profile; get a relationally-coherent fingerprint that survives a getParameter(0x9245) probe. JIT-friendly inject payload, Chromium-native fetch (real Chrome JA4 by definition), behavioral synth on top of Bezier+Fitts. Leaves no crumbs. "
WTF are you talking about? This is incoherent?