OpenBrowser-AI

What it does

OpenBrowser connects AI agents to browser through raw CDP. No abstraction layer. The LLM writes Python in a persistent namespace, batching operations per call. Page state at ~450 characters. Benchmarked against 3 frameworks on 6 real tasks: 100% accuracy across the board, 2.6x fewer tokens, 59% lower inference costs. Methodology is public and reproducible. MIT licensed. CLI + MCP server. 15 LLM providers. Two published RL studies training open-source models for browser control.

Control any browser with natural language. 2-6x fewer tokens than every competitor across MCP and CLI benchmarks. 100% task accuracy. Open source.

Give your AI agent a real browser. 2-6x fewer tokens than every competitor, tested across MCP and CLI benchmarks with 100% accuracy. Tested against 6 competitors across 2 independent benchmarks. Text-first architecture means 2-6x fewer tokens, every time. Works with 15 LLM providers out of the box, plus any OpenAI-compatible endpoint. One execute_code tool, persistent Python namespace. Navigate, click, type, extract. Watch your agent browse in real-time via VNC streaming. See every click, scroll, and navigation live. Docker, Kubernetes, cloud deployment. Battle-tested infrastructure for any scale. MIT licensed. Community-driven. Fully extensible. Build on top of OpenBrowser. The agent…from openbrowser.me

In the maker’s words, at launch

Your AI agent is burning 6x more tokens than it needs to just to browse the web. We built OpenBrowser MCP to fix that. Most browser MCPs give the LLM dozens of tools: click, scroll, type, extract, navigate. Each call dumps the entire page accessibility tree into the context window. One Wikipedia page? 124K+ tokens.…

• 
  Context GatewayMar 2026 · github.com · ▲220

  We built an open-source proxy that sits between coding agents (Claude Code, OpenClaw, etc.) and the LLM, compressing tool outputs before they enter the context window. Demo: https://www.youtube.com/watch?v=-vFZ6MPrwjw#t=9s. Motivation: Agents are terrible at managing context. A single file read or grep can dump thousands of tokens into the window, most of it noise. This isn't just expensive — it actively degrades quality. Long-context benchmarks consistently show steep accuracy drops as context grows (OpenAI's GPT-5.4 eval goes from 97.2% at 32k to 36.6% at 1M…

• 
  Hyperbrowser MCP Server – Connect AI agents to the web through browsers2025 · github.com · ▲63

  Hi HN! Excited to share our MCP Server at Hyperbrowser - something we’ve been working on for a few days. We think it’s a pretty neat way to connect LLMs and IDEs like Cursor / Windsurf to the internet. Our MCP server exposes seven tools for data collection and browsing: 1. `scrape_webpage` - Extract formatted (markdown, screenshot etc) content from any webpage 2. `crawl_webpages` - Navigate through multiple linked pages and extract LLM-friendly formatted content 3. `extract_structured_data` - Convert messy HTML into structured JSON 4. `search_with_bing` - Query the web and get results…

• 
  AI BrowserNov 2025 · aibrowser.surf · ▲143

  Hey HN! I built Browser AI after getting frustrated with the official Playwright MCP server in Claude Code. If you've used it, you know the pain: - 20+ separate tools that Claude has to juggle (click, fill, navigate, wait_for_selector, etc.) - Context window explosions from verbose screenshot data and DOM snapshots - Constant tool selection errors where Claude picks the wrong tool or forgets parameters Browser AI wraps Playwright MCP with a single natural language interface. Instead of Claude managing dozens of tools, you get one: browser_ai.execute. Just tell it what you want in plain…

• 
  I wrote an open-source browser alternative for Computer Use for any LLM2024 · github.com · ▲180

  Hey HN, I made Browser-Use, an open-source tool that lets (all Langchain supported) LLMs execute tasks directly in the browser just with function calling. It allows you to build agents that interact with web elements using natural language prompts. We created a layer that simplifies website interaction for LLMs by extracting xPaths and interactive elements like buttons and input fields (and other fancy things). This enables you to design custom web automation and scraping functions without manual inspection through DevTools. Hasn't this been done a lot of times? Good question, as a general…

• 
  Code ModeNov 2025 · github.com · ▲144

  I’ve been testing a simple idea inspired by Apple/Cloudflare/Anthropic: LLMs are far better at writing a small program than coordinating multiple tool calls. So instead of giving the model 10+ tools, I exposed one: a TypeScript sandbox with access to the same interfaces. The model writes a script → it runs once → done. What changed - +68% reduction in token use - No multi-step drift or retries - Local models (Llama 3.1 8B / Phi-3) became much more reliable Repo: https://github.com/universal-tool-calling-protocol/code-mode Curious whether others have seen…

• 
  Open Browser UseMay 2026 · github.com · ▲112

  Open-source browser automation for local AI agents

• 
  Spira Maxima▲578

  Video model that turns script into viral social media videos

  AI · 17d ago · spira.ai

• 
  Rill Browser▲540

  The browser where Claude Code and Codex work beside you

  AI · 16d ago · rill.love

• 
  Germany's new sovereign AI model Kolibri▲424

  Kolibri is Aleph Alpha

  AI · 18d ago · tej.as

• 
  Giving Opus 5.5 a simulated paint canvas▲392

  Paintings by AI models: programs run through a simulation of oil paint.

  AI · 20d ago · stillwet.art

• 
  Onepin▲384

  AI voiceover with every name, price and date said right

  AI · 30d ago · onepin.ai

• 
  PixVerse R2▲381

  A real-time world model you can explore and change

  AI · 28d ago · world.pixverse.video

• 
  Brila▲1,364

  One-page websites from real Google Maps reviews

  AI · Apr 2026 · brila.ai

• AG
  A game where you build a GPU▲964

  Thought the resources for GPU arch were lacking, so here we are

  Life & fun · Apr 2026 · jaso1024.com

• 
  I built a tiny LLM to demystify how language models work▲915

  Built a ~9M param LLM from scratch to understand how they actually work. Vanilla transformer, 60K synthetic conversations, ~130 lines of PyTorch. Trains in 5 min on a free Colab T4. The fish thinks the meaning of life is food. Fork it and swap the personality for your own character.

  AI · Apr 2026 · github.com

• 
  Fathom 3.0▲800

  AI meeting notes: now bot-free, in ChatGPT & Claude + more

  AI · Apr 2026 · fathom.ai

• 
  Brutalist Concrete Laptop Stand (2024)▲786

  I made a laptop stand out of solid concrete, in the style of brutalist architecture, with themes of urbex and urban decay

  Life & fun · Apr 2026 · sam-burns.com

• 
  I built a frontpage for personal blogs▲783

  With social media and now AI, its important to keep the indie web alive. There are many people who write frequently. Blogosphere tries to highlight them by fetching the recent posts from personal blogs across many categories. There are two versions: Minimal (HN-inspired, fast, static): https://text.blogosphere.app/ Non-minimal: https://blogosphere.app/ If you don't find your blog (or your favorite ones), please add them. I will review and approve it.

  AI · Apr 2026 · text.blogosphere.app