SaaS & Software·Jul 21, 2026

Show HN: Browser Tools SDK – an optimal browser harness for agents

We’re open-sourcing Browser Tools SDK: a small TypeScript package to give any AI agent a reliable way to control a real browser. With just a few lines of code, you can give any agent a production-ready browser harness import { createAiSdkBrowserTools } from "libretto-browser-tools/ai-sdk"; import { LocalBrowserProvider } from "libretto-browser-tools"; const { tools } = createAiSdkBrowserTools(new LocalBrowserProvider()); const result = await generateText({ model: anthropic("claude-sonnet-4-5"), tools, prompt: "Go to Hacker News and tell me the top story", }); We built the Browser Tools SDK because access to a browser is one of the most important tools of a productive agent, but we found many of the existing tools lacking. Using the insights we gained building Libretto, we were able to build the SDK so that it’s significantly more context and cost-efficient than other tools: We compared Browser Tools SDK against agent-browser, playwright-cli, and dev-browser on 26 live-site tasks with GPT 5.6 Sol (best results from 3 runs). Browser Tools tied for the best pass rate at 24/26 and came in about 55% cheaper per successful task than the next alternative ($0.106 vs $0.235) while also using far fewer tokens (1.45M vs 2.29M+). Benchmark methodology is in the repo. The Browser Tools SDK only exposes 6 tools, of which only 2 are really important: ` ` and ` `. The snapshot tool was designed from the ground up to be extremely context-efficient and provide agents the overview they need to know what to try next. The exec tool takes and executes raw Playwright code, of which the models have been trained on heaps of, and know how to use without much direction. Works with AI SDK and Pi out of the box, plus a custom path for anything else. More frameworks are coming soon. You can also use any browser infra provider (Libretto Cloud, Kernel, Browserbase, etc.) or use `LocalBrowserProvider` to run Chromium locally (although that will have less anti-bot protections). MIT licensed. Check out the docs: and join the Discord: . We’d love any and all feedback.

Hacker News2 min readSingle source
Show HN: Browser Tools SDK – an optimal browser harness for agents
Image · Hacker News
The gist
5-point summary · 1 min

We’re open-sourcing Browser Tools SDK: a small TypeScript package to give any AI agent a reliable way to control a real browser. With just a few lines of code, you can give any agent a production-ready browser harness import { createAiSdkBrowserTools } from "libretto-browser-tools/ai-sdk"; import { LocalBrowserProvider } from "libretto-browser-tools"; const { tools } = createAiSdkBrowserTools(new LocalBrowserProvider()); const result = await generateText({ model: anthropic("claude-sonnet-4-5"), tools, prompt: "Go to Hacker News and tell me the top story", }); We built the Browser Tools SDK because access to a browser is one of the most important tools of a productive agent, but we found many of the existing tools lacking. Using the insights we gained building Libretto, we were able to build the SDK so that it’s significantly more context and cost-efficient than other tools: We compared Browser Tools SDK against agent-browser, playwright-cli, and dev-browser on 26 live-site tasks with GPT 5.6 Sol (best results from 3 runs). Browser Tools tied for the best pass rate at 24/26 and came in about 55% cheaper per successful task than the next alternative ($0.106 vs $0.235) while also using far fewer tokens (1.45M vs 2.29M+). Benchmark methodology is in the repo. The Browser Tools SDK only exposes 6 tools, of which only 2 are really important: ` ` and ` `. The snapshot tool was designed from the ground up to be extremely context-efficient and provide agents the overview they need to know what to try next. The exec tool takes and executes raw Playwright code, of which the models have been trained on heaps of, and know how to use without much direction. Works with AI SDK and Pi out of the box, plus a custom path for anything else. More frameworks are coming soon. You can also use any browser infra provider (Libretto Cloud, Kernel, Browserbase, etc.) or use `LocalBrowserProvider` to run Chromium locally (although that will have less anti-bot protections). MIT licensed. Check out the docs: and join the Discord: . We’d love any and all feedback.

  • We’re open-sourcing Browser Tools SDK: a small TypeScript package to give any AI agent a reliable way to control a real browser.
  • Browser Tools tied for the best pass rate at 24/26 and came in about 55% cheaper per successful task than the next alternative ($0.106 vs $0.235) while also using far fewer tokens (1.45M vs 2.29M+).
  • The Browser Tools SDK only exposes 6 tools, of which only 2 are really important: ` ` and ` `.
  • The snapshot tool was designed from the ground up to be extremely context-efficient and provide agents the overview they need to know what to try next.
  • You can also use any browser infra provider (Libretto Cloud, Kernel, Browserbase, etc.) or use `LocalBrowserProvider` to run Chromium locally (although that will have less anti-bot protections).
$0.106$0.23555%

We’re open-sourcing Browser Tools SDK: a small TypeScript package to give any AI agent a reliable way to control a real browser. With just a few lines of code, you can give any agent a production-ready browser harness import { createAiSdkBrowserTools } from "libretto-browser-tools/ai-sdk"; import { LocalBrowserProvider } from "libretto-browser-tools"; const { tools } = createAiSdkBrowserTools(new LocalBrowserProvider()); const result = await generateText({ model: anthropic("claude-sonnet-4-5"), tools, prompt: "Go to Hacker News and tell me the top story", }); We built the Browser Tools SDK because access to a browser is one of the most important tools of a productive agent, but we found many of the existing tools lacking. Using the insights we gained building Libretto, we were able to build the SDK so that it’s significantly more context and cost-efficient than other tools: We compared Browser Tools SDK against agent-browser, playwright-cli, and dev-browser on 26 live-site tasks with GPT 5.6 Sol (best results from 3 runs). Browser Tools tied for the best pass rate at 24/26 and came in about 55% cheaper per successful task than the next alternative ($0.106 vs $0.235) while also using far fewer tokens (1.45M vs 2.29M+). Benchmark methodology is in the repo. The Browser Tools SDK only exposes 6 tools, of which only 2 are really important: ` ` and ` `. The snapshot tool was designed from the ground up to be extremely context-efficient and provide agents the overview they need to know what to try next. The exec tool takes and executes raw Playwright code, of which the models have been trained on heaps of, and know how to use without much direction. Works with AI SDK and Pi out of the box, plus a custom path for anything else. More frameworks are coming soon. You can also use any browser infra provider (Libretto Cloud, Kernel, Browserbase, etc.) or use `LocalBrowserProvider` to run Chromium locally (although that will have less anti-bot protections). MIT licensed. Check out the docs: and join the Discord:. We’d love any and all feedback.

Integrity note  ·  Xela does not rewrite or paraphrase article content. The excerpt above is the source publication's own words, sanitized for display. For the full piece — including any quotes, charts, or images — read it at Hacker News. Xela's rewritten version is off for this story, so there's no editorial angle attached — you're getting the source's reporting unfiltered. When the rewrite is on, we add a What this means block underneath with the operator/trader takeaway.

What people are saying

Discussion

Hot takes

0/280

Loading takes…

Comments

Discussion · 0

Sign in to comment, like, and save articles.

Sign in

Loading comments…

Keep readingSaaS & Software desk
See all in SaaS
Show HN: Computable – Buy, sell, and redeem GPU for the exact weeks you want
·

Show HN: Computable – Buy, sell, and redeem GPU for the exact weeks you want

Hey we are Computable. We spent years building trading infrastructure at Jump Trading and Coinbase. From that point of view, compute looks like energy markets before 2000: everything trades through private bilateral leases, there’s no visible price, and nothing can be resold. The same H100 rents at a 2x spread depending on who’s asking, and once you sign a 24-month lease, it can never change hands. So we built a market for GPU nodes, sold by the calendar week. Here are three things you can do on it that you can’t do anywhere else: - Buy exactly the weeks you need. Three nodes for the last two weeks of October means you pay for those two weeks and walk away when the run is done. You don't have to commit to 6-24 months, and there's no flexibility premium. - Sell back what you don’t use. If your plans change, unused weeks turn back into cash: we keep a market quote posted on every position, so there’s always a price to hit. - Buy (or sell) the future. Lock January in July at a price you set, so the cost of a future run is known before another 40% H100 rate jump makes it not. The first auction is live now at : a block of nodes, August through the end of January. Bidding closes July 31. The bids are sealed. You bid how many nodes, which weeks, and your price. Every clearing price gets published after settlement, because nobody knows what a GPU week is worth and we want the prints to exist in public. Happy to go deep in the comments on anything, including the clearing mechanism, which is a packing problem and was fun to design! Comments URL: Points: 3 # Comments: 1

Hacker NewsSingle source
Newsletter

Track saas & software every morning.

Daily digest tuned to this beat. The 5 stories most worth your time. Unsubscribe anytime.