{"data":{"slug":"magnitudedev-browser-agent","name":"browser-agent","tagline":"Open-source vision-first browser agent for AI automation","github_url":"https://github.com/magnitudedev/browser-agent","owner":"magnitudedev","repo":"browser-agent","owner_avatar_url":"https://avatars.githubusercontent.com/u/199500616?v=4","primary_language":"TypeScript","stars":4128,"forks":238,"topics":["ai","automation","browser","framework","javascipt","playwright","rpa","selenium","test","typescript"],"archived":false,"github_pushed_at":"2026-02-08T20:51:06+00:00","maintenance_label":"Slowing","stars_delta_30d":13,"url":"https://www.graphcanon.com/tools/magnitudedev-browser-agent","markdown_url":"https://www.graphcanon.com/tools/magnitudedev-browser-agent.md","api_url":"https://www.graphcanon.com/api/graphcanon/tools/magnitudedev-browser-agent","graph_url":"https://www.graphcanon.com/api/graphcanon/graph?tool=magnitudedev-browser-agent","description":"Open-source, vision-first browser agent","homepage_url":"https://magnitude.run","license":"Apache-2.0","open_issues":34,"watchers":26,"ai_summary":"Magnitudedev/browser-agent is an open-source project primarily coded in TypeScript that focuses on automating tasks within a web environment using a vision-first approach. It integrates AI capabilities with browser interactions through frameworks like Playwright and can be seen as complementary to RPA solutions such as Selenium.","readme_excerpt":"
\n \"Magnitude\n
\n\n
\n\n
\n \"Documentation\" \"License\" \"Discord\" \"Follow \"Follow\n
\n\n
\n\n> šŸš€ New: Magnitude is state-of-the-art, scoring [94% on WebVoyager](https://github.com/magnitudedev/webvoyager)!\n\nMagnitude uses vision AI to enable you to control your browser with natural language.\n- šŸ§­ **Navigate** - Sees and understands any interface to plan out actions\n- šŸ–±ļø **Interact** - Executes precise actions using mouse and keyboard\n- šŸ” **Extract** - Intelligently extracts useful structured data\n- āœ… **Verify** - Built-in test runner with powerful visual assertions\n\nYou can use it to automate tasks on the web, integrate between apps without APIs, extract data, test your web apps, or as a building block for your own browser agents.\n\n\n\nā†•ļø Magnitude in action! ā†•ļø\n```ts\n// Magnitude can handle high-level tasks\nawait agent.act('Create a task', {\n // Optionally pass data that the agent will use where appropriate\n data: {\n title: 'Use Magnitude',\n description: 'Run \"npx create-magnitude-app\" and follow the instructions',\n },\n});\n\n// It can also handle low-level actions\nawait agent.act('Drag \"Use Magnitude\" to the top of the in progress column');\n\n// Intelligently extract data based on the DOM content matching a provided zod schema\nconst tasks = await agent.extract(\n 'List in progress tasks',\n z.array(z.object({\n title: z.string(),\n description: z.string(),\n // Agent can extract existing data or new insights\n difficulty: z.number().describe('Rate the difficulty between 1-5')\n })),\n);\n```\n\n## Get started\n\n### Running your first browser automation\n```bash\nnpx create-magnitude-app\n```\n\nThis will create a new project and walk you through the steps for setting up Magnitude. It will also create an example script that you can run right away!\n\n### Using the test runner\nTo install the test runner for use in an **existing** web app, please run:\n```bash\nnpm i --save-dev magnitude-test && npx magnitude init\n```\n\nThis will create a basic tests directory `tests/magnitude` with:\n- `magnitude.config.ts`: Magnitude test configuration file\n- `example.mag.ts`: An example test file\n\nFor information on how to run tests and integrate into CI/CD see [here](https://docs.magnitude.run/core-concepts/running-tests).\n\n> [!NOTE]\n> Magnitude requires a large **visually grounded** model. We recommend Claude Sonnet 4 for the best performance, but are also compatible with Qwen-2.5VL 72B. See [docs](https://docs.magnitude.run/customizing/llm-configuration) for more information.\n\n\n## Why Magnitude?\nāŒ **Problem #1:** Most browser agents draw numbered boxes around page elements - doesn't generalize well due to complex modern sites\n\nāœ… **Solution: Vision-first architecture**\n* Visually grounded LLM specifies pixel coordinates\n* True generalization independent of DOM structure\n* Future-proof architecture for desktop apps, VMs, etc.\n\nāŒ **Problem #2:** Most browser agents follow \"high-level prompt + tools =","github_created_at":"2025-03-20T15:50:12+00:00","created_at":"2026-07-15T10:54:27.021185+00:00","updated_at":"2026-09-20T04:49:40.113562+00:00","categories":[{"slug":"ai-agents","name":"AI Agents","url":"https://www.graphcanon.com/categories/ai-agents","markdown_url":"https://www.graphcanon.com/categories/ai-agents.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/ai-agents"},{"slug":"computer-vision","name":"Computer Vision","url":"https://www.graphcanon.com/categories/computer-vision","markdown_url":"https://www.graphcanon.com/categories/computer-vision.md","api_url":"https://www.graphcanon.com/api/graphcanon/categories/computer-vision"}],"tags":[{"slug":"ai","name":"ai"},{"slug":"automation","name":"automation"},{"slug":"browser","name":"browser"},{"slug":"framework","name":"framework"},{"slug":"javascript","name":"javascript"},{"slug":"playwright","name":"playwright"},{"slug":"rpa","name":"rpa"},{"slug":"selenium","name":"selenium"}],"trust":{"provenance":{"is_fork":false,"github_id":952018616,"owner_type":"Organization","methodology":"github_public_v1","parent_repo":null,"near_duplicate_slugs":[]},"computed_at":"2026-09-20T04:49:37.329Z","maintenance":{"label":"Slowing","score":36,"methodology":"github_public_v1","releases_90d":0,"days_since_push":223,"last_release_at":"2026-02-08T20:51:15Z","stars_delta_30d":13,"open_issues_delta_30d":0},"security_summary":{"status":"no_lockfile","scanner":null,"low_count":0,"high_count":0,"last_scan_at":"2026-07-15T10:54:28.224Z","medium_count":0,"scan_profile":"none","critical_count":0}},"capability_facts":{"mcp":{"source":"repo_scan","observed_at":"2026-09-20T04:49:38.581Z","server_manifest":false},"scan":{"source":"repo_scan","observed_at":"2026-09-20T04:49:38.581Z"},"languages":{"value":["typescript","javascript"],"source":"github.language+package.json","observed_at":"2026-09-20T04:49:38.581Z"},"license_spdx":{"value":"Apache-2.0","source":"github.license","observed_at":"2026-09-20T04:49:38.581Z"}},"decision_facts":{"hosting":{"model":"self_hosted","summary":"browser-agent operates as a self-hosted tool; users must download the source code from its repository on Magnitudedev's platform."},"pricing":{"model":"freemium"},"requirements":null,"constraints":{"hosting_model":"self_hosted","pricing_model":"freemium"},"when_to_use":["Use browser-agent if you are working on projects that require vision-centric interactions with websites, as it utilizes a unique vision-first approach for automation tasks.","It is ideal when complementing RPA solutions such as Selenium in scenarios where advanced AI understanding of visual elements enhances task execution."],"when_not_to_use":["Avoid using browser-agent if you need simple text-based web scraping or interactions that do not require sophisticated visual analysis from the web page content.","Do not use if you are looking for a solution solely dedicated to functional testing, as this tool is primarily designed for automating tasks rather than ensuring software functionality."],"source":"enrich:decision_facts","observed_at":"2026-07-16T20:36:59.747Z"},"constraint_facets":{"hosting_model":"self_hosted","pricing_model":"freemium"},"decision_summary":[{"label":"Hosting","value":"self hosted - browser-agent operates as a self-hosted tool; users must download the source code from its repository on Magnitudedev's platform."},{"label":"Pricing","value":"freemium"},{"label":"Adopt for","value":"browser-agent specializes in vision-first web automation tasks using AI capabilities integrated with frameworks like Playwright."},{"label":"License detail","value":"The browser-agent project is licensed under Apache-2.0, allowing free use, modification, and distribution with attribution."}]}}