Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.
JavaScript 3d-generationai-agentsanimationavatardigital-humangenerative-aiglbgltfimage-to-3dllmmcpmcp-servermodel-context-protocolsolanatext-to-3dthreejsweb-componentswebglwebxrx402
feat(hubs): add awesome-browser-agents, a verified 71-entry directory of WebMCP, browser-use, computer-use and headless browser infrastructure for agents
nirholas7acd9ba7d8d2
diff --git a/hubs/awesome-browser-agents/.github/ISSUE_TEMPLATE/add-project.md b/hubs/awesome-browser-agents/.github/ISSUE_TEMPLATE/add-project.mdnew file mode 100644index 000000000..6a73fe3da--- /dev/null+++ b/hubs/awesome-browser-agents/.github/ISSUE_TEMPLATE/add-project.md@@ -0,0 +1,18 @@+---+name: Add a project+about: Suggest a project, standard, or tool for the directory+title: "Add: "+labels: add-project+---++**Name:**++**URL:**++**Repository (if any):**++**Category** (standards, in-page-tools, browser-use-frameworks, browser-mcp-servers, computer-use, automation-foundations, browser-infrastructure, web-data, benchmarks):++**One factual sentence describing it (160 characters max):**++**Status** (active, early, or archived):diff --git a/hubs/awesome-browser-agents/.github/PULL_REQUEST_TEMPLATE.md b/hubs/awesome-browser-agents/.github/PULL_REQUEST_TEMPLATE.mdnew file mode 100644index 000000000..c78e64d40--- /dev/null+++ b/hubs/awesome-browser-agents/.github/PULL_REQUEST_TEMPLATE.md@@ -0,0 +1,16 @@+## What this PR does++- [ ] Adds a new project+- [ ] Updates an existing entry (URL, description, status)++## Project++Name and URL:++## Checklist++- [ ] The project is public and the URL works+- [ ] The description is one factual sentence, 160 characters or fewer+- [ ] The entry is in the right category and sorted alphabetically by name+- [ ] I ran `npm run check` and it passed+- [ ] I committed both `data/entries.json` and the regenerated `README.md`diff --git a/hubs/awesome-browser-agents/CONTRIBUTING.md b/hubs/awesome-browser-agents/CONTRIBUTING.mdnew file mode 100644index 000000000..1a3ce0014--- /dev/null+++ b/hubs/awesome-browser-agents/CONTRIBUTING.md@@ -0,0 +1,61 @@+# Contributing++Thanks for helping map how agents use the web. Contributions of new projects, corrections, and status updates are all welcome.++## What belongs here++A project fits when it is real, public, and relevant to agents operating on or with the web:++- Standards, proposals, and conventions for agent-readable or agent-actionable web content.+- WebMCP and in-page tool libraries; MCP servers and clients for browsers.+- Browser-use and computer-use frameworks, agents, and sandboxes.+- Browser automation engines and headless browser infrastructure used by agents.+- Web data tools built for agents; benchmarks and research agents for web tasks.++Requirements: a working public URL, a factual description, and enough substance to evaluate (docs, a repo, or a published spec). Early projects are welcome with `status: "early"`.++## Adding an entry++Add an object to `data/entries.json`:++```json+{+ "id": "my-project",+ "name": "My Project",+ "url": "https://example.com/",+ "repo": "https://github.com/example/my-project",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "One factual sentence about what it does, at most 160 characters.",+ "status": "active",+ "added": "2026-10-10"+}+```++Field notes:++- `id`: unique kebab-case slug.+- `repo`: source repository URL, or `null` if there is none.+- `category`: an `id` from `data/categories.json`.+- `chain`: always `"none"` for this list.+- `description`: one factual sentence, no marketing language, 160 characters max.+- `status`: `active`, `early`, or `archived`.+- `added`: the date you add the entry, `YYYY-MM-DD`.++Keep entries sorted alphabetically by `name` within each category. Do not use em-dash or en-dash characters; use commas, periods, or hyphens.++Then run:++```sh+npm run check+```++This validates the data and regenerates the tables in `README.md`. Commit both `data/entries.json` and `README.md`.++## Updating or archiving++If a project is renamed, moved, or no longer maintained, open a PR that fixes the URL or sets `status` to `archived`. Archived projects stay listed so people can find them and their lineage.++## Style++Positive, specific, and factual. Describe what a project does, not how it compares to others.diff --git a/hubs/awesome-browser-agents/LICENSE b/hubs/awesome-browser-agents/LICENSEnew file mode 100644index 000000000..0e259d42c--- /dev/null+++ b/hubs/awesome-browser-agents/LICENSE@@ -0,0 +1,121 @@+Creative Commons Legal Code++CC0 1.0 Universal++ CREATIVE COMMONS CORPORATION IS NOT A LAW FIRM AND DOES NOT PROVIDE+ LEGAL SERVICES. DISTRIBUTION OF THIS DOCUMENT DOES NOT CREATE AN+ ATTORNEY-CLIENT RELATIONSHIP. CREATIVE COMMONS PROVIDES THIS+ INFORMATION ON AN "AS-IS" BASIS. CREATIVE COMMONS MAKES NO WARRANTIES+ REGARDING THE USE OF THIS DOCUMENT OR THE INFORMATION OR WORKS+ PROVIDED HEREUNDER, AND DISCLAIMS LIABILITY FOR DAMAGES RESULTING FROM+ THE USE OF THIS DOCUMENT OR THE INFORMATION OR WORKS PROVIDED+ HEREUNDER.++Statement of Purpose++The laws of most jurisdictions throughout the world automatically confer+exclusive Copyright and Related Rights (defined below) upon the creator+and subsequent owner(s) (each and all, an "owner") of an original work of+authorship and/or a database (each, a "Work").++Certain owners wish to permanently relinquish those rights to a Work for+the purpose of contributing to a commons of creative, cultural and+scientific works ("Commons") that the public can reliably and without fear+of later claims of infringement build upon, modify, incorporate in other+works, reuse and redistribute as freely as possible in any form whatsoever+and for any purposes, including without limitation commercial purposes.+These owners may contribute to the Commons to promote the ideal of a free+culture and the further production of creative, cultural and scientific+works, or to gain reputation or greater distribution for their Work in+part through the use and efforts of others.++For these and/or other purposes and motivations, and without any+expectation of additional consideration or compensation, the person+associating CC0 with a Work (the "Affirmer"), to the extent that he or she+is an owner of Copyright and Related Rights in the Work, voluntarily+elects to apply CC0 to the Work and publicly distribute the Work under its+terms, with knowledge of his or her Copyright and Related Rights in the+Work and the meaning and intended legal effect of CC0 on those rights.++1. Copyright and Related Rights. A Work made available under CC0 may be+protected by copyright and related or neighboring rights ("Copyright and+Related Rights"). Copyright and Related Rights include, but are not+limited to, the following:++ i. the right to reproduce, adapt, distribute, perform, display,+ communicate, and translate a Work;+ ii. moral rights retained by the original author(s) and/or performer(s);+iii. publicity and privacy rights pertaining to a person's image or+ likeness depicted in a Work;+ iv. rights protecting against unfair competition in regards to a Work,+ subject to the limitations in paragraph 4(a), below;+ v. rights protecting the extraction, dissemination, use and reuse of data+ in a Work;+ vi. database rights (such as those arising under Directive 96/9/EC of the+ European Parliament and of the Council of 11 March 1996 on the legal+ protection of databases, and under any national implementation+ thereof, including any amended or successor version of such+ directive); and+vii. other similar, equivalent or corresponding rights throughout the+ world based on applicable law or treaty, and any national+ implementations thereof.++2. Waiver. To the greatest extent permitted by, but not in contravention+of, applicable law, Affirmer hereby overtly, fully, permanently,+irrevocably and unconditionally waives, abandons, and surrenders all of+Affirmer's Copyright and Related Rights and associated claims and causes+of action, whether now known or unknown (including existing as well as+future claims and causes of action), in the Work (i) in all territories+worldwide, (ii) for the maximum duration provided by applicable law or+treaty (including future time extensions), (iii) in any current or future+medium and for any number of copies, and (iv) for any purpose whatsoever,+including without limitation commercial, advertising or promotional+purposes (the "Waiver"). Affirmer makes the Waiver for the benefit of each+member of the public at large and to the detriment of Affirmer's heirs and+successors, fully intending that such Waiver shall not be subject to+revocation, rescission, cancellation, termination, or any other legal or+equitable action to disrupt the quiet enjoyment of the Work by the public+as contemplated by Affirmer's express Statement of Purpose.++3. Public License Fallback. Should any part of the Waiver for any reason+be judged legally invalid or ineffective under applicable law, then the+Waiver shall be preserved to the maximum extent permitted taking into+account Affirmer's express Statement of Purpose. In addition, to the+extent the Waiver is so judged Affirmer hereby grants to each affected+person a royalty-free, non transferable, non sublicensable, non exclusive,+irrevocable and unconditional license to exercise Affirmer's Copyright and+Related Rights in the Work (i) in all territories worldwide, (ii) for the+maximum duration provided by applicable law or treaty (including future+time extensions), (iii) in any current or future medium and for any number+of copies, and (iv) for any purpose whatsoever, including without+limitation commercial, advertising or promotional purposes (the+"License"). The License shall be deemed effective as of the date CC0 was+applied by Affirmer to the Work. Should any part of the License for any+reason be judged legally invalid or ineffective under applicable law, such+partial invalidity or ineffectiveness shall not invalidate the remainder+of the License, and in such case Affirmer hereby affirms that he or she+will not (i) exercise any of his or her remaining Copyright and Related+Rights in the Work or (ii) assert any associated claims and causes of+action with respect to the Work, in either case contrary to Affirmer's+express Statement of Purpose.++4. Limitations and Disclaimers.++ a. No trademark or patent rights held by Affirmer are waived, abandoned,+ surrendered, licensed or otherwise affected by this document.+ b. Affirmer offers the Work as-is and makes no representations or+ warranties of any kind concerning the Work, express, implied,+ statutory or otherwise, including without limitation warranties of+ title, merchantability, fitness for a particular purpose, non+ infringement, or the absence of latent or other defects, accuracy, or+ the present or absence of errors, whether or not discoverable, all to+ the greatest extent permissible under applicable law.+ c. Affirmer disclaims responsibility for clearing rights of other persons+ that may apply to the Work or any use thereof, including without+ limitation any person's Copyright and Related Rights in the Work.+ Further, Affirmer disclaims responsibility for obtaining any necessary+ consents, permissions or other rights required for any use of the+ Work.+ d. Affirmer understands and acknowledges that Creative Commons is not a+ party to this document and has no duty or obligation with respect to+ this CC0 or use of the Work.diff --git a/hubs/awesome-browser-agents/README.md b/hubs/awesome-browser-agents/README.mdnew file mode 100644index 000000000..ca07118e5--- /dev/null+++ b/hubs/awesome-browser-agents/README.md@@ -0,0 +1,179 @@+# Awesome Browser Agents++A curated, community-fed directory of the projects, standards, and infrastructure that let AI agents use the web: WebMCP and in-page tool standards, browser-use and computer-use frameworks, agent-readable web conventions such as `llms.txt`, MCP servers for browsers, and headless browser infrastructure.++The web is becoming a place where agents work alongside people. This list exists so that builders can find the right building block quickly, and so that every good project in the space has one neutral, well-organized place to be discovered.++Every entry lives in [`data/entries.json`](data/entries.json), so the list is machine-readable as well as human-readable. The tables below are generated from that file.++## Add your project++Open a pull request that adds one object to [`data/entries.json`](data/entries.json). The short version:++1. Copy an existing entry and edit it. Keep entries sorted alphabetically by name within their category.+2. Run `npm run check` (Node 18+, no dependencies). It validates your entry and regenerates this README.+3. Open the PR using the template. See [CONTRIBUTING.md](CONTRIBUTING.md) for the inclusion bar and field reference.++Prefer not to edit JSON? [Open an "Add a project" issue](.github/ISSUE_TEMPLATE/add-project.md) and a maintainer will add it.++## Machine-readable data++- [`data/entries.json`](data/entries.json): all entries.+- [`data/categories.json`](data/categories.json): ordered categories.+- [`data/schema.json`](data/schema.json): JSON Schema for entries.++## Projects++<!-- ENTRIES:START -->++**71 projects** across 9 categories.++- [Standards and Proposals](#standards-and-proposals) (11)+- [WebMCP and In-Page Agent Tools](#webmcp-and-in-page-agent-tools) (6)+- [Browser-Use Frameworks](#browser-use-frameworks) (12)+- [Browser MCP Servers](#browser-mcp-servers) (8)+- [Computer-Use Agents](#computer-use-agents) (13)+- [Automation Foundations](#automation-foundations) (5)+- [Headless Browser Infrastructure](#headless-browser-infrastructure) (6)+- [Web Data for Agents](#web-data-for-agents) (4)+- [Benchmarks and Research](#benchmarks-and-research) (6)++### Standards and Proposals++Specifications, protocols, and conventions that let agents discover, read, and act on the web.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [A2A Protocol](https://a2a-protocol.org/) | Agent2Agent, an open protocol for communication and interoperability between opaque agentic applications. | [repo](https://github.com/a2aproject/A2A) | active |+| [Agent Skills](https://agentskills.io/) | An open format for packaging reusable skills that agents can discover and load on demand. | - | active |+| [AGENTS.md](https://agents.md/) | A simple, open format of project instructions that gives coding agents context about a repository. | - | active |+| [ai.robots.txt](https://github.com/ai-robots-txt/ai.robots.txt) | A maintained list of AI agents and crawlers, with ready-made robots.txt rules for sites that want to manage them. | [repo](https://github.com/ai-robots-txt/ai.robots.txt) | active |+| [llms.txt](https://llmstxt.org/) | A proposal from Jeremy Howard to publish a Markdown file at a site root that gives LLMs a curated guide to its content. | [repo](https://github.com/AnswerDotAI/llms-txt) | active |+| [Model Context Protocol](https://modelcontextprotocol.io/) | The open protocol for connecting AI applications to tools, data sources, and workflows. | [repo](https://github.com/modelcontextprotocol/modelcontextprotocol) | active |+| [NLWeb](https://github.com/nlweb-ai/NLWeb) | The reference implementation of NLWeb, which lets sites offer natural language interfaces to their own content. | [repo](https://github.com/nlweb-ai/NLWeb) | active |+| [Robots Exclusion Protocol (RFC 9309)](https://www.rfc-editor.org/info/rfc9309/) | The IETF standard that formalizes robots.txt rules for crawlers and automated clients. | - | active |+| [Web Bot Auth](https://www.webbotauth.org/) | A neutral directory and signing scheme for authenticated bots, built on signed HTTP requests. | [repo](https://github.com/cloudflare/web-bot-auth) | early |+| [WebMCP in Chrome](https://developer.chrome.com/docs/ai/webmcp) | Chrome documentation for WebMCP, covering imperative and declarative tools and the origin trial. | - | early |+| [WebMCP Specification](https://webmachinelearning.github.io/webmcp/) | A draft Web Machine Learning Community Group API that lets web apps offer JavaScript tools to AI agents. | [repo](https://github.com/webmachinelearning/webmcp) | early |++### WebMCP and In-Page Agent Tools++Libraries and demos that let a web page expose tools to agents, and MCP running in or alongside the browser.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Early WebMCP Proposal](https://github.com/jasonjmcghee/WebMCP) | An early WebMCP proposal and implementation that later fed into the W3C community group work. | [repo](https://github.com/jasonjmcghee/WebMCP) | archived |+| [MCP TypeScript SDK](https://github.com/modelcontextprotocol/typescript-sdk) | The official TypeScript SDK for building Model Context Protocol servers and clients. | [repo](https://github.com/modelcontextprotocol/typescript-sdk) | active |+| [MCP-B](https://mcp-b.ai/) | Open-source WebMCP packages and a browser extension for publishing and calling tools on the web. | [repo](https://github.com/WebMCP-org/npm-packages) | active |+| [mcp-use](https://github.com/mcp-use/mcp-use) | A full-stack MCP framework for building MCP apps for ChatGPT and Claude and MCP servers for agents. | [repo](https://github.com/mcp-use/mcp-use) | active |+| [WebLLM](https://github.com/mlc-ai/web-llm) | A high-performance in-browser LLM inference engine for running models locally inside a web page. | [repo](https://github.com/mlc-ai/web-llm) | active |+| [WebMCP Tools](https://github.com/GoogleChromeLabs/webmcp-tools) | Developer utilities and demos from Chrome Labs that support adoption of the WebMCP API. | [repo](https://github.com/GoogleChromeLabs/webmcp-tools) | active |++### Browser-Use Frameworks++Open-source libraries and SDKs that let LLMs drive a browser to complete tasks.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [agent-browser](https://agent-browser.dev/) | A browser automation CLI built for AI agents. | [repo](https://github.com/vercel-labs/agent-browser) | active |+| [AgentQL](https://www.agentql.com/) | A query language and Playwright integration for locating elements and extracting data from any page. | [repo](https://github.com/tinyfish-io/agentql) | active |+| [Browser Use](https://browser-use.com/) | An open-source library that lets AI agents use the browser. | [repo](https://github.com/browser-use/browser-use) | active |+| [Browser Use Web UI](https://github.com/browser-use/web-ui) | A web interface for running Browser Use agents in your browser. | [repo](https://github.com/browser-use/web-ui) | active |+| [HyperAgent](https://github.com/hyperbrowserai/HyperAgent) | An AI browser automation framework from the Hyperbrowser team. | [repo](https://github.com/hyperbrowserai/HyperAgent) | active |+| [LaVague](https://github.com/lavague-ai/LaVague) | A Large Action Model framework for developing AI web agents. | [repo](https://github.com/lavague-ai/LaVague) | active |+| [Midscene.js](https://github.com/web-infra-dev/midscene) | A vision-driven framework that lets you control web and other interfaces in natural language. | [repo](https://github.com/web-infra-dev/midscene) | active |+| [Nanobrowser](https://github.com/nanobrowser/nanobrowser) | An open-source Chrome extension that runs multi-agent web automation workflows with your own LLM API key. | [repo](https://github.com/nanobrowser/nanobrowser) | active |+| [Open Operator](https://github.com/browserbase/open-operator) | A template for building web agents with Stagehand on Browserbase. | [repo](https://github.com/browserbase/open-operator) | archived |+| [Skyvern](https://www.skyvern.com/) | Automates browser-based workflows with AI. | [repo](https://github.com/Skyvern-AI/skyvern) | active |+| [Stagehand](https://www.stagehand.dev/) | An SDK for browser agents to extract data and interact with any site on the web. | [repo](https://github.com/browserbase/stagehand) | active |+| [Workflow Use](https://github.com/browser-use/workflow-use) | Create and run repeatable browser workflows, described by its authors as RPA 2.0. | [repo](https://github.com/browser-use/workflow-use) | early |++### Browser MCP Servers++MCP servers that give assistants control of, or insight into, a real browser.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Browser MCP](https://browsermcp.io/) | An MCP server and extension that lets AI applications control your own browser. | [repo](https://github.com/BrowserMCP/mcp) | active |+| [Browser Tools MCP](https://github.com/AgentDeskAI/browser-tools-mcp) | An MCP server and Chrome extension that surfaces browser logs to Cursor and other MCP-compatible IDEs. | [repo](https://github.com/AgentDeskAI/browser-tools-mcp) | active |+| [Browserbase MCP Server](https://github.com/browserbase/mcp-server-browserbase) | An MCP server that lets LLMs control a browser with Browserbase and Stagehand. | [repo](https://github.com/browserbase/mcp-server-browserbase) | archived |+| [Chrome DevTools MCP](https://github.com/ChromeDevTools/chrome-devtools-mcp) | Chrome DevTools for coding agents, exposed as an MCP server. | [repo](https://github.com/ChromeDevTools/chrome-devtools-mcp) | active |+| [Firecrawl MCP Server](https://github.com/firecrawl/firecrawl-mcp-server) | The official Firecrawl MCP server, adding web scraping and search to Cursor, Claude, and other clients. | [repo](https://github.com/firecrawl/firecrawl-mcp-server) | active |+| [mcp-chrome](https://github.com/hangwin/mcp-chrome) | A Chrome extension based MCP server that exposes browser functionality to AI assistants. | [repo](https://github.com/hangwin/mcp-chrome) | active |+| [Playwright MCP](https://github.com/microsoft/playwright-mcp) | The Playwright MCP server from Microsoft, giving agents structured browser control. | [repo](https://github.com/microsoft/playwright-mcp) | active |+| [Playwright MCP Server (ExecuteAutomation)](https://github.com/executeautomation/mcp-playwright) | A Playwright MCP server for automating browsers and APIs in Claude Desktop, Cline, Cursor, and more. | [repo](https://github.com/executeautomation/mcp-playwright) | active |++### Computer-Use Agents++Agents, sandboxes, and tooling that operate full graphical desktops from screenshots and actions.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Agent S](https://github.com/simular-ai/Agent-S) | An open agentic framework that uses computers like a human. | [repo](https://github.com/simular-ai/Agent-S) | active |+| [All-in-One Sandbox](https://github.com/agent-infra/sandbox) | An all-in-one sandbox for AI agents that combines browser and related tools in a single environment. | [repo](https://github.com/agent-infra/sandbox) | active |+| [Anthropic Computer Use Tool](https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool) | Claude API documentation for the computer use tool that controls a desktop through screenshots and actions. | - | active |+| [Claude Quickstarts](https://github.com/anthropics/claude-quickstarts) | A collection of starter projects for building deployable applications with the Claude API. | [repo](https://github.com/anthropics/claude-quickstarts) | active |+| [Cua](https://cua.ai/) | Open-source drivers, cross-OS fleets, and benchmarks for computer-use agents. | [repo](https://github.com/trycua/cua) | active |+| [E2B Desktop](https://github.com/e2b-dev/desktop) | An E2B sandbox with a desktop graphical environment that you can connect to any LLM for secure computer use. | [repo](https://github.com/e2b-dev/desktop) | active |+| [Gemini Computer Use](https://ai.google.dev/gemini-api/docs/computer-use) | Gemini API documentation for the computer use capability that lets models operate user interfaces. | - | active |+| [Magentic-UI](https://github.com/microsoft/magentic-ui) | An experimental Microsoft agent that works across the browser and the local file system. | [repo](https://github.com/microsoft/magentic-ui) | early |+| [OmniParser](https://github.com/microsoft/OmniParser) | A screen parsing tool for pure vision based GUI agents. | [repo](https://github.com/microsoft/OmniParser) | active |+| [OpenAI Computer Use Guide](https://developers.openai.com/api/docs/guides/tools-computer-use) | OpenAI API documentation for the computer use tool. | - | active |+| [OpenAI CUA Sample App](https://github.com/openai/openai-cua-sample-app) | Shows how to use OpenAI Computer Using Agent via the API across multiple computer environments. | [repo](https://github.com/openai/openai-cua-sample-app) | active |+| [Surf](https://github.com/e2b-dev/surf) | A computer-use agent powered by OpenAI that drives an E2B virtual desktop through natural language. | [repo](https://github.com/e2b-dev/surf) | active |+| [UI-TARS Desktop](https://github.com/bytedance/UI-TARS-desktop) | An open-source multimodal AI agent stack that connects cutting-edge models with agent infrastructure. | [repo](https://github.com/bytedance/UI-TARS-desktop) | active |++### Automation Foundations++The browser automation engines and headless browsers that agent frameworks build on.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Chrome DevTools](https://developer.chrome.com/docs/devtools/) | The documentation hub for Chrome DevTools, the debugging surface that agent tooling builds on. | - | active |+| [Lightpanda](https://lightpanda.io/) | A headless browser designed for AI and automation. | [repo](https://github.com/lightpanda-io/browser) | active |+| [Playwright](https://playwright.dev/) | A framework for web testing and automation across Chromium, Firefox, and WebKit with a single API. | [repo](https://github.com/microsoft/playwright) | active |+| [Puppeteer](https://pptr.dev/) | A JavaScript API for controlling Chrome and Firefox. | [repo](https://github.com/puppeteer/puppeteer) | active |+| [Puppeteer for Cloudflare](https://github.com/cloudflare/puppeteer) | A Puppeteer Core fork that works with Cloudflare Browser Workers. | [repo](https://github.com/cloudflare/puppeteer) | active |++### Headless Browser Infrastructure++Self-hostable and hosted browser fleets built for agent workloads.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Anchor Browser](https://anchorbrowser.io/) | Secure browser infrastructure for computer-use agents. | - | active |+| [Browserbase](https://www.browserbase.com/) | Cloud headless browsers for AI agents and the company behind Stagehand. | - | active |+| [Browserless](https://www.browserless.io/) | Headless browsers deployable in Docker or in the cloud, reachable over MCP, Puppeteer, or Playwright. | [repo](https://github.com/browserless/browserless) | active |+| [Hyperbrowser](https://www.hyperbrowser.ai/) | Cloud browsers for AI agents and apps. | - | active |+| [Kernel](https://www.kernel.sh/) | Browser infrastructure for web agents and automations. | - | active |+| [Steel Browser](https://steel.dev/) | An open-source browser API and sandbox for AI agents and apps. | [repo](https://github.com/steel-dev/steel-browser) | active |++### Web Data for Agents++Crawlers, readers, and extraction tools that turn pages into agent-ready content.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [Crawl4AI](https://crawl4ai.com/) | An open-source crawler and scraper that turns websites into clean, LLM-ready Markdown. | [repo](https://github.com/unclecode/crawl4ai) | active |+| [Crawlee](https://github.com/apify/crawlee) | A web scraping and browser automation library for Node.js that works with Puppeteer, Playwright, Cheerio, and JSDOM. | [repo](https://github.com/apify/crawlee) | active |+| [Firecrawl](https://www.firecrawl.dev/) | A web data API that gives AI agents clean data from the web. | [repo](https://github.com/firecrawl/firecrawl) | active |+| [Jina Reader](https://github.com/jina-ai/reader) | Converts any URL into LLM-friendly input by prefixing it with r.jina.ai. | [repo](https://github.com/jina-ai/reader) | active |++### Benchmarks and Research++Datasets, environments, and research agents for measuring and improving web agents.++| Project | Description | Repo | Status |+| --- | --- | --- | --- |+| [AutoWebGLM](https://github.com/THUDM/AutoWebGLM) | An LLM-based web navigating agent from KDD 2024. | [repo](https://github.com/THUDM/AutoWebGLM) | active |+| [BrowserGym](https://github.com/ServiceNow/BrowserGym) | A Gym environment for web task automation. | [repo](https://github.com/ServiceNow/BrowserGym) | active |+| [Mind2Web](https://osu-nlp-group.github.io/Mind2Web/) | A NeurIPS 2023 dataset and benchmark for generalist agents for the web. | [repo](https://github.com/OSU-NLP-Group/Mind2Web) | active |+| [SeeAct](https://github.com/OSU-NLP-Group/SeeAct) | An ICML 2024 system for generalist web agents that complete tasks on any website using large multimodal models. | [repo](https://github.com/OSU-NLP-Group/SeeAct) | active |+| [WebArena](https://webarena.dev/) | A realistic web environment and suite of benchmarks for building autonomous web agents. | [repo](https://github.com/web-arena-x/webarena) | active |+| [WebVoyager](https://github.com/MinorJerry/WebVoyager) | Code for WebVoyager, an end-to-end web agent built with large multimodal models. | [repo](https://github.com/MinorJerry/WebVoyager) | active |++<!-- ENTRIES:END -->++## License++[CC0 1.0 Universal](LICENSE). The list is dedicated to the public domain: use it, fork it, and build on it freely. Listed projects keep their own licenses and trademarks.diff --git a/hubs/awesome-browser-agents/data/categories.json b/hubs/awesome-browser-agents/data/categories.jsonnew file mode 100644index 000000000..89b507c87--- /dev/null+++ b/hubs/awesome-browser-agents/data/categories.json@@ -0,0 +1,11 @@+[+ {"id": "standards", "title": "Standards and Proposals", "blurb": "Specifications, protocols, and conventions that let agents discover, read, and act on the web."},+ {"id": "in-page-tools", "title": "WebMCP and In-Page Agent Tools", "blurb": "Libraries and demos that let a web page expose tools to agents, and MCP running in or alongside the browser."},+ {"id": "browser-use-frameworks", "title": "Browser-Use Frameworks", "blurb": "Open-source libraries and SDKs that let LLMs drive a browser to complete tasks."},+ {"id": "browser-mcp-servers", "title": "Browser MCP Servers", "blurb": "MCP servers that give assistants control of, or insight into, a real browser."},+ {"id": "computer-use", "title": "Computer-Use Agents", "blurb": "Agents, sandboxes, and tooling that operate full graphical desktops from screenshots and actions."},+ {"id": "automation-foundations", "title": "Automation Foundations", "blurb": "The browser automation engines and headless browsers that agent frameworks build on."},+ {"id": "browser-infrastructure", "title": "Headless Browser Infrastructure", "blurb": "Self-hostable and hosted browser fleets built for agent workloads."},+ {"id": "web-data", "title": "Web Data for Agents", "blurb": "Crawlers, readers, and extraction tools that turn pages into agent-ready content."},+ {"id": "benchmarks", "title": "Benchmarks and Research", "blurb": "Datasets, environments, and research agents for measuring and improving web agents."}+]diff --git a/hubs/awesome-browser-agents/data/entries.json b/hubs/awesome-browser-agents/data/entries.jsonnew file mode 100644index 000000000..2b77f21cd--- /dev/null+++ b/hubs/awesome-browser-agents/data/entries.json@@ -0,0 +1,783 @@+[+ {+ "id": "a2a-protocol",+ "name": "A2A Protocol",+ "url": "https://a2a-protocol.org/",+ "repo": "https://github.com/a2aproject/A2A",+ "category": "standards",+ "chain": "none",+ "description": "Agent2Agent, an open protocol for communication and interoperability between opaque agentic applications.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "agent-skills",+ "name": "Agent Skills",+ "url": "https://agentskills.io/",+ "repo": null,+ "category": "standards",+ "chain": "none",+ "description": "An open format for packaging reusable skills that agents can discover and load on demand.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "agents-md",+ "name": "AGENTS.md",+ "url": "https://agents.md/",+ "repo": null,+ "category": "standards",+ "chain": "none",+ "description": "A simple, open format of project instructions that gives coding agents context about a repository.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "ai-robots-txt",+ "name": "ai.robots.txt",+ "url": "https://github.com/ai-robots-txt/ai.robots.txt",+ "repo": "https://github.com/ai-robots-txt/ai.robots.txt",+ "category": "standards",+ "chain": "none",+ "description": "A maintained list of AI agents and crawlers, with ready-made robots.txt rules for sites that want to manage them.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "llms-txt",+ "name": "llms.txt",+ "url": "https://llmstxt.org/",+ "repo": "https://github.com/AnswerDotAI/llms-txt",+ "category": "standards",+ "chain": "none",+ "description": "A proposal from Jeremy Howard to publish a Markdown file at a site root that gives LLMs a curated guide to its content.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "model-context-protocol",+ "name": "Model Context Protocol",+ "url": "https://modelcontextprotocol.io/",+ "repo": "https://github.com/modelcontextprotocol/modelcontextprotocol",+ "category": "standards",+ "chain": "none",+ "description": "The open protocol for connecting AI applications to tools, data sources, and workflows.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "nlweb",+ "name": "NLWeb",+ "url": "https://github.com/nlweb-ai/NLWeb",+ "repo": "https://github.com/nlweb-ai/NLWeb",+ "category": "standards",+ "chain": "none",+ "description": "The reference implementation of NLWeb, which lets sites offer natural language interfaces to their own content.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "robots-exclusion-protocol-rfc-9309",+ "name": "Robots Exclusion Protocol (RFC 9309)",+ "url": "https://www.rfc-editor.org/info/rfc9309/",+ "repo": null,+ "category": "standards",+ "chain": "none",+ "description": "The IETF standard that formalizes robots.txt rules for crawlers and automated clients.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "web-bot-auth",+ "name": "Web Bot Auth",+ "url": "https://www.webbotauth.org/",+ "repo": "https://github.com/cloudflare/web-bot-auth",+ "category": "standards",+ "chain": "none",+ "description": "A neutral directory and signing scheme for authenticated bots, built on signed HTTP requests.",+ "status": "early",+ "added": "2026-10-10"+ },+ {+ "id": "webmcp-in-chrome",+ "name": "WebMCP in Chrome",+ "url": "https://developer.chrome.com/docs/ai/webmcp",+ "repo": null,+ "category": "standards",+ "chain": "none",+ "description": "Chrome documentation for WebMCP, covering imperative and declarative tools and the origin trial.",+ "status": "early",+ "added": "2026-10-10"+ },+ {+ "id": "webmcp-specification",+ "name": "WebMCP Specification",+ "url": "https://webmachinelearning.github.io/webmcp/",+ "repo": "https://github.com/webmachinelearning/webmcp",+ "category": "standards",+ "chain": "none",+ "description": "A draft Web Machine Learning Community Group API that lets web apps offer JavaScript tools to AI agents.",+ "status": "early",+ "added": "2026-10-10"+ },+ {+ "id": "early-webmcp-proposal",+ "name": "Early WebMCP Proposal",+ "url": "https://github.com/jasonjmcghee/WebMCP",+ "repo": "https://github.com/jasonjmcghee/WebMCP",+ "category": "in-page-tools",+ "chain": "none",+ "description": "An early WebMCP proposal and implementation that later fed into the W3C community group work.",+ "status": "archived",+ "added": "2026-10-10"+ },+ {+ "id": "mcp-typescript-sdk",+ "name": "MCP TypeScript SDK",+ "url": "https://github.com/modelcontextprotocol/typescript-sdk",+ "repo": "https://github.com/modelcontextprotocol/typescript-sdk",+ "category": "in-page-tools",+ "chain": "none",+ "description": "The official TypeScript SDK for building Model Context Protocol servers and clients.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "mcp-b",+ "name": "MCP-B",+ "url": "https://mcp-b.ai/",+ "repo": "https://github.com/WebMCP-org/npm-packages",+ "category": "in-page-tools",+ "chain": "none",+ "description": "Open-source WebMCP packages and a browser extension for publishing and calling tools on the web.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "mcp-use",+ "name": "mcp-use",+ "url": "https://github.com/mcp-use/mcp-use",+ "repo": "https://github.com/mcp-use/mcp-use",+ "category": "in-page-tools",+ "chain": "none",+ "description": "A full-stack MCP framework for building MCP apps for ChatGPT and Claude and MCP servers for agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "webllm",+ "name": "WebLLM",+ "url": "https://github.com/mlc-ai/web-llm",+ "repo": "https://github.com/mlc-ai/web-llm",+ "category": "in-page-tools",+ "chain": "none",+ "description": "A high-performance in-browser LLM inference engine for running models locally inside a web page.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "webmcp-tools",+ "name": "WebMCP Tools",+ "url": "https://github.com/GoogleChromeLabs/webmcp-tools",+ "repo": "https://github.com/GoogleChromeLabs/webmcp-tools",+ "category": "in-page-tools",+ "chain": "none",+ "description": "Developer utilities and demos from Chrome Labs that support adoption of the WebMCP API.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "agent-browser",+ "name": "agent-browser",+ "url": "https://agent-browser.dev/",+ "repo": "https://github.com/vercel-labs/agent-browser",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A browser automation CLI built for AI agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "agentql",+ "name": "AgentQL",+ "url": "https://www.agentql.com/",+ "repo": "https://github.com/tinyfish-io/agentql",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A query language and Playwright integration for locating elements and extracting data from any page.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browser-use",+ "name": "Browser Use",+ "url": "https://browser-use.com/",+ "repo": "https://github.com/browser-use/browser-use",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "An open-source library that lets AI agents use the browser.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browser-use-web-ui",+ "name": "Browser Use Web UI",+ "url": "https://github.com/browser-use/web-ui",+ "repo": "https://github.com/browser-use/web-ui",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A web interface for running Browser Use agents in your browser.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "hyperagent",+ "name": "HyperAgent",+ "url": "https://github.com/hyperbrowserai/HyperAgent",+ "repo": "https://github.com/hyperbrowserai/HyperAgent",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "An AI browser automation framework from the Hyperbrowser team.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "lavague",+ "name": "LaVague",+ "url": "https://github.com/lavague-ai/LaVague",+ "repo": "https://github.com/lavague-ai/LaVague",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A Large Action Model framework for developing AI web agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "midscene-js",+ "name": "Midscene.js",+ "url": "https://github.com/web-infra-dev/midscene",+ "repo": "https://github.com/web-infra-dev/midscene",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A vision-driven framework that lets you control web and other interfaces in natural language.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "nanobrowser",+ "name": "Nanobrowser",+ "url": "https://github.com/nanobrowser/nanobrowser",+ "repo": "https://github.com/nanobrowser/nanobrowser",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "An open-source Chrome extension that runs multi-agent web automation workflows with your own LLM API key.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "open-operator",+ "name": "Open Operator",+ "url": "https://github.com/browserbase/open-operator",+ "repo": "https://github.com/browserbase/open-operator",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "A template for building web agents with Stagehand on Browserbase.",+ "status": "archived",+ "added": "2026-10-10"+ },+ {+ "id": "skyvern",+ "name": "Skyvern",+ "url": "https://www.skyvern.com/",+ "repo": "https://github.com/Skyvern-AI/skyvern",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "Automates browser-based workflows with AI.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "stagehand",+ "name": "Stagehand",+ "url": "https://www.stagehand.dev/",+ "repo": "https://github.com/browserbase/stagehand",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "An SDK for browser agents to extract data and interact with any site on the web.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "workflow-use",+ "name": "Workflow Use",+ "url": "https://github.com/browser-use/workflow-use",+ "repo": "https://github.com/browser-use/workflow-use",+ "category": "browser-use-frameworks",+ "chain": "none",+ "description": "Create and run repeatable browser workflows, described by its authors as RPA 2.0.",+ "status": "early",+ "added": "2026-10-10"+ },+ {+ "id": "browser-mcp",+ "name": "Browser MCP",+ "url": "https://browsermcp.io/",+ "repo": "https://github.com/BrowserMCP/mcp",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "An MCP server and extension that lets AI applications control your own browser.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browser-tools-mcp",+ "name": "Browser Tools MCP",+ "url": "https://github.com/AgentDeskAI/browser-tools-mcp",+ "repo": "https://github.com/AgentDeskAI/browser-tools-mcp",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "An MCP server and Chrome extension that surfaces browser logs to Cursor and other MCP-compatible IDEs.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browserbase-mcp-server",+ "name": "Browserbase MCP Server",+ "url": "https://github.com/browserbase/mcp-server-browserbase",+ "repo": "https://github.com/browserbase/mcp-server-browserbase",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "An MCP server that lets LLMs control a browser with Browserbase and Stagehand.",+ "status": "archived",+ "added": "2026-10-10"+ },+ {+ "id": "chrome-devtools-mcp",+ "name": "Chrome DevTools MCP",+ "url": "https://github.com/ChromeDevTools/chrome-devtools-mcp",+ "repo": "https://github.com/ChromeDevTools/chrome-devtools-mcp",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "Chrome DevTools for coding agents, exposed as an MCP server.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "firecrawl-mcp-server",+ "name": "Firecrawl MCP Server",+ "url": "https://github.com/firecrawl/firecrawl-mcp-server",+ "repo": "https://github.com/firecrawl/firecrawl-mcp-server",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "The official Firecrawl MCP server, adding web scraping and search to Cursor, Claude, and other clients.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "mcp-chrome",+ "name": "mcp-chrome",+ "url": "https://github.com/hangwin/mcp-chrome",+ "repo": "https://github.com/hangwin/mcp-chrome",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "A Chrome extension based MCP server that exposes browser functionality to AI assistants.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "playwright-mcp",+ "name": "Playwright MCP",+ "url": "https://github.com/microsoft/playwright-mcp",+ "repo": "https://github.com/microsoft/playwright-mcp",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "The Playwright MCP server from Microsoft, giving agents structured browser control.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "playwright-mcp-server-executeautomation",+ "name": "Playwright MCP Server (ExecuteAutomation)",+ "url": "https://github.com/executeautomation/mcp-playwright",+ "repo": "https://github.com/executeautomation/mcp-playwright",+ "category": "browser-mcp-servers",+ "chain": "none",+ "description": "A Playwright MCP server for automating browsers and APIs in Claude Desktop, Cline, Cursor, and more.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "agent-s",+ "name": "Agent S",+ "url": "https://github.com/simular-ai/Agent-S",+ "repo": "https://github.com/simular-ai/Agent-S",+ "category": "computer-use",+ "chain": "none",+ "description": "An open agentic framework that uses computers like a human.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "all-in-one-sandbox",+ "name": "All-in-One Sandbox",+ "url": "https://github.com/agent-infra/sandbox",+ "repo": "https://github.com/agent-infra/sandbox",+ "category": "computer-use",+ "chain": "none",+ "description": "An all-in-one sandbox for AI agents that combines browser and related tools in a single environment.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "anthropic-computer-use-tool",+ "name": "Anthropic Computer Use Tool",+ "url": "https://platform.claude.com/docs/en/agents-and-tools/tool-use/computer-use-tool",+ "repo": null,+ "category": "computer-use",+ "chain": "none",+ "description": "Claude API documentation for the computer use tool that controls a desktop through screenshots and actions.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "claude-quickstarts",+ "name": "Claude Quickstarts",+ "url": "https://github.com/anthropics/claude-quickstarts",+ "repo": "https://github.com/anthropics/claude-quickstarts",+ "category": "computer-use",+ "chain": "none",+ "description": "A collection of starter projects for building deployable applications with the Claude API.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "cua",+ "name": "Cua",+ "url": "https://cua.ai/",+ "repo": "https://github.com/trycua/cua",+ "category": "computer-use",+ "chain": "none",+ "description": "Open-source drivers, cross-OS fleets, and benchmarks for computer-use agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "e2b-desktop",+ "name": "E2B Desktop",+ "url": "https://github.com/e2b-dev/desktop",+ "repo": "https://github.com/e2b-dev/desktop",+ "category": "computer-use",+ "chain": "none",+ "description": "An E2B sandbox with a desktop graphical environment that you can connect to any LLM for secure computer use.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "gemini-computer-use",+ "name": "Gemini Computer Use",+ "url": "https://ai.google.dev/gemini-api/docs/computer-use",+ "repo": null,+ "category": "computer-use",+ "chain": "none",+ "description": "Gemini API documentation for the computer use capability that lets models operate user interfaces.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "magentic-ui",+ "name": "Magentic-UI",+ "url": "https://github.com/microsoft/magentic-ui",+ "repo": "https://github.com/microsoft/magentic-ui",+ "category": "computer-use",+ "chain": "none",+ "description": "An experimental Microsoft agent that works across the browser and the local file system.",+ "status": "early",+ "added": "2026-10-10"+ },+ {+ "id": "omniparser",+ "name": "OmniParser",+ "url": "https://github.com/microsoft/OmniParser",+ "repo": "https://github.com/microsoft/OmniParser",+ "category": "computer-use",+ "chain": "none",+ "description": "A screen parsing tool for pure vision based GUI agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "openai-computer-use-guide",+ "name": "OpenAI Computer Use Guide",+ "url": "https://developers.openai.com/api/docs/guides/tools-computer-use",+ "repo": null,+ "category": "computer-use",+ "chain": "none",+ "description": "OpenAI API documentation for the computer use tool.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "openai-cua-sample-app",+ "name": "OpenAI CUA Sample App",+ "url": "https://github.com/openai/openai-cua-sample-app",+ "repo": "https://github.com/openai/openai-cua-sample-app",+ "category": "computer-use",+ "chain": "none",+ "description": "Shows how to use OpenAI Computer Using Agent via the API across multiple computer environments.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "surf",+ "name": "Surf",+ "url": "https://github.com/e2b-dev/surf",+ "repo": "https://github.com/e2b-dev/surf",+ "category": "computer-use",+ "chain": "none",+ "description": "A computer-use agent powered by OpenAI that drives an E2B virtual desktop through natural language.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "ui-tars-desktop",+ "name": "UI-TARS Desktop",+ "url": "https://github.com/bytedance/UI-TARS-desktop",+ "repo": "https://github.com/bytedance/UI-TARS-desktop",+ "category": "computer-use",+ "chain": "none",+ "description": "An open-source multimodal AI agent stack that connects cutting-edge models with agent infrastructure.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "chrome-devtools",+ "name": "Chrome DevTools",+ "url": "https://developer.chrome.com/docs/devtools/",+ "repo": null,+ "category": "automation-foundations",+ "chain": "none",+ "description": "The documentation hub for Chrome DevTools, the debugging surface that agent tooling builds on.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "lightpanda",+ "name": "Lightpanda",+ "url": "https://lightpanda.io/",+ "repo": "https://github.com/lightpanda-io/browser",+ "category": "automation-foundations",+ "chain": "none",+ "description": "A headless browser designed for AI and automation.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "playwright",+ "name": "Playwright",+ "url": "https://playwright.dev/",+ "repo": "https://github.com/microsoft/playwright",+ "category": "automation-foundations",+ "chain": "none",+ "description": "A framework for web testing and automation across Chromium, Firefox, and WebKit with a single API.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "puppeteer",+ "name": "Puppeteer",+ "url": "https://pptr.dev/",+ "repo": "https://github.com/puppeteer/puppeteer",+ "category": "automation-foundations",+ "chain": "none",+ "description": "A JavaScript API for controlling Chrome and Firefox.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "puppeteer-for-cloudflare",+ "name": "Puppeteer for Cloudflare",+ "url": "https://github.com/cloudflare/puppeteer",+ "repo": "https://github.com/cloudflare/puppeteer",+ "category": "automation-foundations",+ "chain": "none",+ "description": "A Puppeteer Core fork that works with Cloudflare Browser Workers.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "anchor-browser",+ "name": "Anchor Browser",+ "url": "https://anchorbrowser.io/",+ "repo": null,+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "Secure browser infrastructure for computer-use agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browserbase",+ "name": "Browserbase",+ "url": "https://www.browserbase.com/",+ "repo": null,+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "Cloud headless browsers for AI agents and the company behind Stagehand.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browserless",+ "name": "Browserless",+ "url": "https://www.browserless.io/",+ "repo": "https://github.com/browserless/browserless",+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "Headless browsers deployable in Docker or in the cloud, reachable over MCP, Puppeteer, or Playwright.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "hyperbrowser",+ "name": "Hyperbrowser",+ "url": "https://www.hyperbrowser.ai/",+ "repo": null,+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "Cloud browsers for AI agents and apps.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "kernel",+ "name": "Kernel",+ "url": "https://www.kernel.sh/",+ "repo": null,+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "Browser infrastructure for web agents and automations.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "steel-browser",+ "name": "Steel Browser",+ "url": "https://steel.dev/",+ "repo": "https://github.com/steel-dev/steel-browser",+ "category": "browser-infrastructure",+ "chain": "none",+ "description": "An open-source browser API and sandbox for AI agents and apps.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "crawl4ai",+ "name": "Crawl4AI",+ "url": "https://crawl4ai.com/",+ "repo": "https://github.com/unclecode/crawl4ai",+ "category": "web-data",+ "chain": "none",+ "description": "An open-source crawler and scraper that turns websites into clean, LLM-ready Markdown.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "crawlee",+ "name": "Crawlee",+ "url": "https://github.com/apify/crawlee",+ "repo": "https://github.com/apify/crawlee",+ "category": "web-data",+ "chain": "none",+ "description": "A web scraping and browser automation library for Node.js that works with Puppeteer, Playwright, Cheerio, and JSDOM.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "firecrawl",+ "name": "Firecrawl",+ "url": "https://www.firecrawl.dev/",+ "repo": "https://github.com/firecrawl/firecrawl",+ "category": "web-data",+ "chain": "none",+ "description": "A web data API that gives AI agents clean data from the web.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "jina-reader",+ "name": "Jina Reader",+ "url": "https://github.com/jina-ai/reader",+ "repo": "https://github.com/jina-ai/reader",+ "category": "web-data",+ "chain": "none",+ "description": "Converts any URL into LLM-friendly input by prefixing it with r.jina.ai.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "autowebglm",+ "name": "AutoWebGLM",+ "url": "https://github.com/THUDM/AutoWebGLM",+ "repo": "https://github.com/THUDM/AutoWebGLM",+ "category": "benchmarks",+ "chain": "none",+ "description": "An LLM-based web navigating agent from KDD 2024.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "browsergym",+ "name": "BrowserGym",+ "url": "https://github.com/ServiceNow/BrowserGym",+ "repo": "https://github.com/ServiceNow/BrowserGym",+ "category": "benchmarks",+ "chain": "none",+ "description": "A Gym environment for web task automation.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "mind2web",+ "name": "Mind2Web",+ "url": "https://osu-nlp-group.github.io/Mind2Web/",+ "repo": "https://github.com/OSU-NLP-Group/Mind2Web",+ "category": "benchmarks",+ "chain": "none",+ "description": "A NeurIPS 2023 dataset and benchmark for generalist agents for the web.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "seeact",+ "name": "SeeAct",+ "url": "https://github.com/OSU-NLP-Group/SeeAct",+ "repo": "https://github.com/OSU-NLP-Group/SeeAct",+ "category": "benchmarks",+ "chain": "none",+ "description": "An ICML 2024 system for generalist web agents that complete tasks on any website using large multimodal models.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "webarena",+ "name": "WebArena",+ "url": "https://webarena.dev/",+ "repo": "https://github.com/web-arena-x/webarena",+ "category": "benchmarks",+ "chain": "none",+ "description": "A realistic web environment and suite of benchmarks for building autonomous web agents.",+ "status": "active",+ "added": "2026-10-10"+ },+ {+ "id": "webvoyager",+ "name": "WebVoyager",+ "url": "https://github.com/MinorJerry/WebVoyager",+ "repo": "https://github.com/MinorJerry/WebVoyager",+ "category": "benchmarks",+ "chain": "none",+ "description": "Code for WebVoyager, an end-to-end web agent built with large multimodal models.",+ "status": "active",+ "added": "2026-10-10"+ }+]diff --git a/hubs/awesome-browser-agents/data/schema.json b/hubs/awesome-browser-agents/data/schema.jsonnew file mode 100644index 000000000..a2f18317b--- /dev/null+++ b/hubs/awesome-browser-agents/data/schema.json@@ -0,0 +1,21 @@+{+ "$schema": "http://json-schema.org/draft-07/schema#",+ "title": "awesome-browser-agents entries",+ "type": "array",+ "items": {+ "type": "object",+ "additionalProperties": false,+ "required": ["id", "name", "url", "repo", "category", "chain", "description", "status", "added"],+ "properties": {+ "id": {"type": "string", "pattern": "^[a-z0-9]+(-[a-z0-9]+)*$"},+ "name": {"type": "string", "minLength": 1},+ "url": {"type": "string", "format": "uri", "pattern": "^https?://"},+ "repo": {"type": ["string", "null"], "pattern": "^https?://"},+ "category": {"type": "string", "description": "An id from data/categories.json"},+ "chain": {"type": "string", "enum": ["none"]},+ "description": {"type": "string", "minLength": 1, "maxLength": 160},+ "status": {"type": "string", "enum": ["active", "early", "archived"]},+ "added": {"type": "string", "pattern": "^\\d{4}-\\d{2}-\\d{2}$"}+ }+ }+}diff --git a/hubs/awesome-browser-agents/package.json b/hubs/awesome-browser-agents/package.jsonnew file mode 100644index 000000000..3130095d3--- /dev/null+++ b/hubs/awesome-browser-agents/package.json@@ -0,0 +1,14 @@+{+ "name": "awesome-browser-agents",+ "version": "1.0.0",+ "private": true,+ "description": "A curated, PR-fed, machine-readable directory of web-native and browser agent projects.",+ "license": "CC0-1.0",+ "type": "module",+ "scripts": {+ "validate": "node scripts/validate.mjs",+ "build": "node scripts/build-readme.mjs",+ "check": "node scripts/validate.mjs && node scripts/build-readme.mjs"+ },+ "engines": {"node": ">=18"}+}diff --git a/hubs/awesome-browser-agents/scripts/build-readme.mjs b/hubs/awesome-browser-agents/scripts/build-readme.mjsnew file mode 100644index 000000000..0dc8552f0--- /dev/null+++ b/hubs/awesome-browser-agents/scripts/build-readme.mjs@@ -0,0 +1,46 @@+#!/usr/bin/env node+// Regenerates the entries section of README.md from data/*.json. No dependencies.+import { readFileSync, writeFileSync } from 'node:fs';+import { fileURLToPath } from 'node:url';+import { dirname, join } from 'node:path';++const root = join(dirname(fileURLToPath(import.meta.url)), '..');+const load = (f) => JSON.parse(readFileSync(join(root, 'data', f), 'utf8'));+const entries = load('entries.json');+const categories = load('categories.json');++const START = '<!-- ENTRIES:START -->';+const END = '<!-- ENTRIES:END -->';+const esc = (s) => s.replace(/\|/g, '\\|');+const anchor = (t) => t.toLowerCase().replace(/[^a-z0-9 -]/g, '').trim().replace(/ /g, '-');+const badge = { active: 'active', early: 'early', archived: 'archived' };++const parts = [];+for (const c of categories) {+ const rows = entries.filter((e) => e.category === c.id);+ if (!rows.length) continue;+ parts.push(`### ${c.title}\n\n${c.blurb}\n`);+ parts.push('| Project | Description | Repo | Status |\n| --- | --- | --- | --- |');+ for (const e of rows) {+ const repo = e.repo ? `[repo](${e.repo})` : '-';+ parts.push(`| [${esc(e.name)}](${e.url}) | ${esc(e.description)} | ${repo} | ${badge[e.status]} |`);+ }+ parts.push('');+}+const index = categories+ .filter((c) => entries.some((e) => e.category === c.id))+ .map((c) => `- [${c.title}](#${anchor(c.title)}) (${entries.filter((e) => e.category === c.id).length})`)+ .join('\n');++const body = `${START}\n\n**${entries.length} projects** across ${categories.length} categories.\n\n${index}\n\n${parts.join('\n')}\n${END}`;++const path = join(root, 'README.md');+const readme = readFileSync(path, 'utf8');+const s = readme.indexOf(START);+const e = readme.indexOf(END);+if (s === -1 || e === -1 || e < s) {+ console.error(`README.md must contain ${START} and ${END} markers.`);+ process.exit(1);+}+writeFileSync(path, readme.slice(0, s) + body + readme.slice(e + END.length));+console.log(`README.md updated with ${entries.length} entries.`);diff --git a/hubs/awesome-browser-agents/scripts/validate.mjs b/hubs/awesome-browser-agents/scripts/validate.mjsnew file mode 100644index 000000000..8477d3a4a--- /dev/null+++ b/hubs/awesome-browser-agents/scripts/validate.mjs@@ -0,0 +1,78 @@+#!/usr/bin/env node+// Validates data/entries.json against data/categories.json. No dependencies.+import { readFileSync } from 'node:fs';+import { fileURLToPath } from 'node:url';+import { dirname, join } from 'node:path';++const root = join(dirname(fileURLToPath(import.meta.url)), '..');+const load = (f) => JSON.parse(readFileSync(join(root, 'data', f), 'utf8'));++const errors = [];+const err = (m) => errors.push(m);++let entries, categories;+try {+ entries = load('entries.json');+ categories = load('categories.json');+} catch (e) {+ console.error(`Could not read data files: ${e.message}`);+ process.exit(1);+}++const DASHES = /[\u2013\u2014]/;+const STATUS = ['active', 'early', 'archived'];+const FIELDS = ['id', 'name', 'url', 'repo', 'category', 'chain', 'description', 'status', 'added'];+const catIds = new Set(categories.map((c) => c.id));+const seen = new Set();++function validUrl(u) {+ try {+ const p = new URL(u);+ return p.protocol === 'https:' || p.protocol === 'http:';+ } catch {+ return false;+ }+}++if (!Array.isArray(entries)) err('entries.json must be an array');+else {+ entries.forEach((e, i) => {+ const tag = `entry #${i + 1} (${e && e.id ? e.id : 'no id'})`;+ for (const f of FIELDS) if (!(f in e)) err(`${tag}: missing field "${f}"`);+ for (const k of Object.keys(e)) if (!FIELDS.includes(k)) err(`${tag}: unknown field "${k}"`);+ if (typeof e.id !== 'string' || !/^[a-z0-9]+(-[a-z0-9]+)*$/.test(e.id)) err(`${tag}: id must be kebab-case`);+ else if (seen.has(e.id)) err(`${tag}: duplicate id`);+ else seen.add(e.id);+ if (typeof e.name !== 'string' || !e.name.trim()) err(`${tag}: name must be a non-empty string`);+ if (!validUrl(e.url)) err(`${tag}: url is not a valid http(s) URL`);+ if (e.repo !== null && !validUrl(e.repo)) err(`${tag}: repo must be a valid URL or null`);+ if (!catIds.has(e.category)) err(`${tag}: category "${e.category}" is not in categories.json`);+ if (e.chain !== 'none') err(`${tag}: chain must be "none"`);+ if (typeof e.description !== 'string' || !e.description.trim()) err(`${tag}: description is required`);+ else if (e.description.length > 160) err(`${tag}: description is ${e.description.length} chars (max 160)`);+ if (!STATUS.includes(e.status)) err(`${tag}: status must be one of ${STATUS.join(', ')}`);+ if (typeof e.added !== 'string' || !/^\d{4}-\d{2}-\d{2}$/.test(e.added) || Number.isNaN(Date.parse(e.added)))+ err(`${tag}: added must be a valid YYYY-MM-DD date`);+ if (DASHES.test(JSON.stringify(e))) err(`${tag}: contains an em-dash or en-dash; use a comma, period, or hyphen`);+ });++ const lastByCat = new Map();+ for (const e of entries) {+ const prev = lastByCat.get(e.category);+ if (prev && e.name.localeCompare(prev, 'en', { sensitivity: 'base' }) < 0)+ err(`"${e.name}" is out of order in category "${e.category}" (comes after "${prev}"); sort alphabetically by name`);+ lastByCat.set(e.category, e.name);+ }+}++for (const c of categories) {+ if (!c.id || !c.title || !c.blurb) err(`category ${JSON.stringify(c.id)} needs id, title, and blurb`);+ if (DASHES.test(JSON.stringify(c))) err(`category "${c.id}" contains an em-dash or en-dash`);+}++if (errors.length) {+ console.error(`Validation failed with ${errors.length} problem${errors.length === 1 ? '' : 's'}:\n`);+ for (const m of errors) console.error(` - ${m}`);+ process.exit(1);+}+console.log(`OK: ${entries.length} entries across ${categories.length} categories.`);