Browser4 – an open-source browser engine for agents and concurrency
Hi HN, I’d like to share an open-source project we’ve been working on for a while: Browser4. The motivation came from a recurring frustration: most browser automation tools (Playwright, Selenium, Puppeteer) are excellent for human-written scripts, but start to show friction when used as a core execution layer for AI agents or at very high concurrency. So instead of building “another wrapper around Playwright”, we experimented with a different direction: designing a browser engine where AI agents are first-class citizens. ### What Browser4 is Browser4 is a browser automation engine built on…
What it does
In the maker’s words, at launch
Hi HN, I’d like to share an open-source project we’ve been working on for a while: Browser4. The motivation came from a recurring frustration: most browser automation tools (Playwright, Selenium, Puppeteer) are excellent for human-written scripts, but start to show friction when used as a core execution layer for AI agents or at very high concurrency. So instead of building “another wrapper around Playwright”, we experimented with a different direction: designing a browser engine where AI agents are first-class citizens. ### What Browser4 is Browser4 is a browser automation engine built on native Chrome DevTools Protocol (CDP), with a focus on: * Coroutine-safe concurrency (designed to run many browser sessions in parallel) * Agent-oriented APIs (navigation, interaction, extraction as composable actions) * Hybrid extraction: ML agent driven extraction + LLM extraction + structured selectors + an SQL-like DOM query language (X-SQL) * Low-level control without Playwright-style abstraction overhead It’s written in Kotlin/JVM, mainly because we needed predictable concurrency behavior and long-running stability under load. The project is fully open-source (Apache 2.0). ### What it’s not * It’s not a drop-in Playwright replacement. * It’s not a no-code RPA tool. * It’s not “LLM magic” — LLMs sit outside the browser engine. Browser4 intentionally stays close to the browser execution layer and leaves planning/reasoning to external agent loops. ### Current use cases we’re testing * Large-scale web data extraction * Agentic workflows (search → navigate → extract → summarize) * Price / content monitoring with frequent revisits * High-concurrency crawling where browser startup and context switching are bottlenecks On a single machine, we can sustain very high daily page visits, though we’re still validating benchmarks across different workloads. ### Open questions (where I’d love feedback) * For agentic systems, does it make sense to bypass Playwright entirely and work closer to CDP? * Where do you see the biggest pain points when combining LLMs with browser automation today? * Is JVM a reasonable choice here, or is Python still the better tradeoff despite concurrency limits * What abstractions would you want in a browser engine built for AI agents? ### Links * GitHub: https://github.com/platonai/browser4 * Website (light overview): https://browser4.io Happy to answer technical questions or hear criticism — especially from people running browser automation or agent systems in production. Thanks for reading.
Does the same job
all alternatives →
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, December 2025
the whole month →- GP
Life & fun · Dec 2025 · dosaygo-studio.github.io
- JG
Hi everyone! My name's Luke and I made the original Jmail here alongside Riley Walz. We had a ton of friends collaborate on building out more of the app suite last night in lieue of DOJ's "Epstein files" release. Please AMA!
Life & fun · Dec 2025 · jmail.world
PlanEat AI▲733AI turns your health goals into a 7-day menu & grocery list
AI · Dec 2025 · planeatai.com
- 2G
Community, All the HN belong to you. This is an archive of hacker news that fits in your browser. When I made HN Made of Primes I realized I could probably do this offline sqlite/wasm thing with the whole GBs of archive. The whole dataset. So I tried it, and this is it. Have Hacker News on your device. Go to this repo (https://github.com/DOSAYGO-STUDIO/HackerBook): you can download it. Big Query -> ETL -> npx serve docs - that's it. 20 years of HN arguments and beauty, can be yours forever. So they'll never die. Ever. It's the unkillable static archive of HN and it's…
Dev tools · Dec 2025 · hackerbook.dosaygo.com
