Hacker News Reader

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

hmm interesting - sometimes it definitely fails yes. Will take a look!

btw - our biggest challenge is exactly this, solving thousands of issues that arise on the fly.

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

What’s your take - how can we expose Browser Use to as many use cases as possible? Is there easier way than openapi config?

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

imo if the website doesn't want us there the long term value is anyway not great (maybe exception is SERP apis or sth which live exlusively because google search api is brutally expensive).

> extract data behind login walls

We mean this more from a perspective of companies wanting it, but there is a login wall. For example (actual customer) - "I am a compliance company that has system from 2001 and interacting with it really painful. Let's use Browser Use to use the search bar, download data and report back to me".

I believe in the long run agents will have to pay for the data from website providers, and then the incentives are once again aligned.

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

We are experimenting with this. Currently the library api is very raw but technically possible (we introduced this notion of initial actions, which are just deterministic actions before the LLM kicks in) - https://github.com/browser-use/browser-use/blob/main/example....

The other way to achieve this with Browser Use is to save the history from `history = agent.run()` and rerun it with `agent.rerun_history(history)`.

I'd love to see if this can of any use to you!

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

I think our cloud is much simpler (just one prompt and go). But it's also sort of a different service. The main differences come from the open source side - we are essentially building more of a framework for anytime to use and they are just a web app.

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

how would that work? Can you control the browser without debug mode? Especially in production the browsers are anyway running on single instance docker containers so the file system is not accesible... are there exploits that can do harm from a virtual machine?

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

hmm, I though about this a lot. But tbh I think MCP is sort of a gimmick... probably the better way is for agents just to understand the http apis directly. Maybe I'm wrong, very happy to be convinced differently. Do you think MCP server for the cloud version would be useful?

gregpr07on Launch HN: Browser Use (YC W25) – open-source web agents

Yeah, we are much bigger and work on a higher level. stagehand work step by step, we are trying to make end to end web agents.

gregpr07on Show HN: Mastra – Open-source JS agent framework, by the developers of Gatsby

Any timeline for python?

gregpr07on Introducing Operator

https://github.com/browser-use/browser-use :)

gregpr07on Operator research preview

If you are worried about costs you can use Browser Use with deepseek which becomes super cheap! https://github.com/browser-use/browser-use

gregpr07on Introducing Operator

It's not even SOTA. The actual SOTA is Browser Use (report here https://browser-use.com/posts/sota-technical-report)

gregpr07on Introducing Operator

Yeah, and Browser Use already has 89% on WebVoyager https://browser-use.com/posts/sota-technical-report

gregpr07on Show HN: Hyperbrowser – Scalable Browser Infrastructure for AI Apps

Creator of Browser Use here. This looks super cool! How do you spin them up so fast, firecracker?

gregpr07on Show HN: Steel.dev – An open-source browser API for AI agents and apps

Nice!! We are working on a higher level library over at https://github.com/gregpr07/browser-use.

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Haha with browser use or Computer use??

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Technically you can yeah, I am not sure if the performance is the same - would have to test it!

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Who (or still in stealth haha)?

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

For that we wanted to give you more control as well with a for loop and you take actions step by step. I think all of these crewai like agent swarms are also very much black boxes.

How would you imagine the perfect scenario? What would make LLM outputs less of a black box?

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Do you think they do any super fancy magic other than for example how ferret ui does their classification of ui elements? It could be very interesting to test head to head hope much better you can make computer use by adding html (it’s much better from our quick testing, just don’t know the numbers).

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Nope: 1280x1024 low resolution with gpt-4o are 85 tokens so approx $0.0002 (so 100x cheaper). For high resolution its apporx $0.002 https://openai.com/api/pricing/

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Hmm, but this how we handle it? We just have a CLI that outputs exactly, goal, state, and asks user for more clarity if needed, no GUI. The original idea was to make it completely headless.

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Could you elaborate on the CLI idea? I am intrigued but not exactly sure what you mean.

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

I don’t know a lot about this but do you have full power of Selenium or not? That would be also very interesting aproach especially when “local” browser models get very good

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Will def try it.

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Fair, also Claude probably only gets better on this since they kinda want people to use Computer use. We are gonna try to do best of both worlds.

Thanks man, Magnus came up with it this morning haha!

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Thanks man, starred yours too, it's super cool to see all these projects getting spun up!

I see Cerebellum is vision only. Did you try adding HTML + screenshot? I think that improves the performance like crazy and you don't have to use Claude only.

Just saw Skyvern today on previous Show HNs haha :)

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Looks nice. I find the cleaning HTML step in our cleaning pipeline extremely important, otherwise there is no real benefit from just using a general vision model and clicking coordinates (and whole HTML is just way too many tokens). How do you guys handle that?

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

Thanks! Have you tried captcha solving with [1]? It's very tricky sometimes, especially with non standard "verify human" - maybe you could solve it by writing Selenium/Javascript code directly and then execute it.

gregpr07on Show HN: I wrote an open-source browser alternative for Computer Use for any LLM

You can run browser use in your terminal, no need for Docker containers. Just clone it and run it