{"id":26896934,"url":"https://github.com/nottelabs/notte","last_synced_at":"2026-03-11T15:01:22.644Z","repository":{"id":267389094,"uuid":"900152988","full_name":"nottelabs/notte","owner":"nottelabs","description":"🌸 Best framework to build web agents, and deploy serverless web automation functions on reliable browser infra.","archived":false,"fork":false,"pushed_at":"2026-03-03T18:27:15.000Z","size":69344,"stargazers_count":1875,"open_issues_count":15,"forks_count":168,"subscribers_count":14,"default_branch":"main","last_synced_at":"2026-03-03T20:04:32.782Z","etag":null,"topics":["agent","ai","anthropic","automation","browser","llm","openai","web"],"latest_commit_sha":null,"homepage":"https://www.notte.cc/","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"other","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/nottelabs.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":"docs/CONTRIBUTING.md","funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null,"zenodo":null,"notice":null,"maintainers":null,"copyright":"COPYRIGHT.md","agents":null,"dco":null,"cla":null}},"created_at":"2024-12-08T02:06:30.000Z","updated_at":"2026-03-03T18:27:28.000Z","dependencies_parsed_at":null,"dependency_job_id":"529587a0-359b-460c-9173-a3e4e7bcdcb5","html_url":"https://github.com/nottelabs/notte","commit_stats":null,"previous_names":["nottelabs/notte","nottelabs/notte-operator"],"tags_count":60,"template":false,"template_full_name":null,"purl":"pkg:github/nottelabs/notte","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nottelabs%2Fnotte","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nottelabs%2Fnotte/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nottelabs%2Fnotte/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nottelabs%2Fnotte/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/nottelabs","download_url":"https://codeload.github.com/nottelabs/notte/tar.gz/refs/heads/main","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/nottelabs%2Fnotte/sbom","scorecard":null,"host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":286080680,"owners_count":30385018,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2026-03-11T14:10:17.325Z","status":"ssl_error","status_checked_at":"2026-03-11T14:09:37.934Z","response_time":84,"last_error":"SSL_connect returned=1 errno=0 peeraddr=140.82.121.6:443 state=error: unexpected eof while reading","robots_txt_status":"success","robots_txt_updated_at":"2025-07-24T06:49:26.215Z","robots_txt_url":"https://github.com/robots.txt","online":false,"can_crawl_api":true,"host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agent","ai","anthropic","automation","browser","llm","openai","web"],"created_at":"2025-04-01T04:02:24.070Z","updated_at":"2026-03-11T15:01:22.637Z","avatar_url":"https://github.com/nottelabs.png","language":"Python","funding_links":[],"categories":["Browser Automation","Python","Browser-extensions","Projects","Frameworks","Text","Task and Project Management","Web Automation and UI Interaction","Agent Categories","Web \u0026 Search","Tools","Browser \u0026 Web Automation"],"sub_categories":["AI Automation","Frameworks \u0026 Models","Advanced Components","Developer tools","Browser Automation","\u003ca name=\"Unclassified\"\u003e\u003c/a\u003eUnclassified","Agent Web Interaction"],"readme":"# Rapidly build reliable web automation agents\n\n\u003cdiv align=\"center\"\u003e\n  \u003cp\u003e\n    The web agent framework built for \u003cstrong\u003espeed\u003c/strong\u003e, \u003cstrong\u003ecost-efficiency\u003c/strong\u003e, \u003cstrong\u003escale\u003c/strong\u003e, and \u003cstrong\u003ereliability\u003c/strong\u003e \u003cbr/\u003e\n    → Read more at: \u003ca href=\"https://github.com/nottelabs/open-operator-evals\" target=\"_blank\" rel=\"noopener noreferrer\"\u003eopen-operator-evals\u003c/a\u003e • \u003ca href=\"https://x.com/nottecore?ref=github\" target=\"_blank\" rel=\"noopener noreferrer\"\u003eX\u003c/a\u003e • \u003ca href=\"https://www.linkedin.com/company/nottelabsinc/?ref=github\" target=\"_blank\" rel=\"noopener noreferrer\"\u003eLinkedIn\u003c/a\u003e • \u003ca href=\"https://notte.cc?ref=github\" target=\"_blank\" rel=\"noopener noreferrer\"\u003eLanding\u003c/a\u003e • \u003ca href=\"https://console.notte.cc/?ref=github\" target=\"_blank\" rel=\"noopener noreferrer\"\u003eConsole\u003c/a\u003e\n  \u003c/p\u003e\n\u003c/div\u003e\n\n\u003cp align=\"center\"\u003e\n  \u003cimg src=\"docs/logo/bgd.png\" alt=\"Notte Logo\" width=\"100%\"\u003e\n\u003c/p\u003e\n\n[![GitHub stars](https://img.shields.io/github/stars/nottelabs/notte?style=social)](https://github.com/nottelabs/notte/stargazers)\n[![License: SSPL-1.0](https://img.shields.io/badge/License-SSPL%201.0-blue.svg)](https://spdx.org/licenses/SSPL-1.0.html)\n[![Python 3.11+](https://img.shields.io/badge/python-3.11+-blue.svg)](https://www.python.org/downloads/)\n[![PyPI version](https://img.shields.io/pypi/v/notte?color=blue)](https://pypi.org/project/notte/)\n[![PyPI Downloads](https://static.pepy.tech/badge/notte?color=blue)](https://pepy.tech/projects/notte)\n\n---\n\n# What is Notte?\n\nNotte provides all the essential tools for building and deploying AI agents that interact seamlessly with the web. Our full-stack framework combines AI agents with traditional scripting for maximum efficiency - letting you script deterministic parts and use AI only when needed, cutting costs by 50%+ while improving reliability. We allow you to develop, deploy, and scale your own agents and web automations, all with a single API. Read more in our documentation [here](https://docs.notte.cc) 🔥\n\n**Opensource Core:**\n- **[Run web agents](#using-python-sdk-recommended)** → Give AI agents natural language tasks to complete on websites\n- **[Structured Output](#structured-output)** → Get data in your exact format with Pydantic models\n- **[Site Interactions](#scraping)** → Observe website states, scrape data and execute actions using Playwright compatible primitives and natural language commands\n\n**API service (Recommended)**\n- **[Stealth Browser Sessions](#session-features)** → Browser instances with built-in CAPTCHA solving, proxies, and anti-detection\n- **[Hybrid Workflows](#workflows)** → Combine scripting and AI agents to reduce costs and improve reliability\n- **[Secrets Vaults](#agent-vault)** → Enterprise-grade credential management to store emails, passwords, MFA tokens, SSO, etc.\n- **[Digital Personas](#agent-persona)** → Create digital identities with unique emails, phones, and automated 2FA for account creation workflows\n\n# Quickstart\n\n```\npip install notte\npatchright install --with-deps chromium\n```\n\n### Run in local mode\n\nUse the following script to spinup an agent using opensource features (you'll need your own LLM API keys):\n\n```python\nimport notte\nfrom dotenv import load_dotenv\nload_dotenv()\n\nwith notte.Session(headless=False) as session:\n    agent = notte.Agent(session=session, reasoning_model='gemini/gemini-2.5-flash', max_steps=30)\n    response = agent.run(task=\"doom scroll cat memes on google images\")\n```\n\n### Using Python SDK (Recommended)\n\nWe also provide an effortless API that hosts the browser sessions for you - and provide plenty of premium features. To run the agent you'll need to first sign up on the [Notte Console](https://console.notte.cc) and create a free Notte API key 🔑\n\n```python\nfrom notte_sdk import NotteClient\nimport os\n\nclient = NotteClient(api_key=os.getenv(\"NOTTE_API_KEY\"))\n\nwith client.Session(open_viewer=True) as session:\n    agent = client.Agent(session=session, reasoning_model='gemini/gemini-2.5-flash', max_steps=30)\n    response = agent.run(task=\"doom scroll cat memes on google images\")\n```\n\nOur setup allows you to experiment locally, then drop-in replace the import and prefix `notte` objects with `cli` to switch to SDK and get hosted browser sessions plus access to premium features!\n\n# Benchmarks\n\n| Rank | Provider                                                    | Agent Self-Report | LLM Evaluation | Time per Task | Task Reliability |\n| ---- | ----------------------------------------------------------- | ----------------- | -------------- | ------------- | ---------------- |\n| 🏆   | [Notte](https://github.com/nottelabs/notte)                 | **86.2%**         | **79.0%**      | **47s**       | **96.6%**        |\n| 2️⃣   | [Browser-Use](https://github.com/browser-use/browser-use)   | 77.3%             | 60.2%          | 113s          | 83.3%            |\n| 3️⃣   | [Convergence](https://github.com/convergence-ai/proxy-lite) | 38.4%             | 31.4%          | 83s           | 50%              |\n\nRead the full story here: [https://github.com/nottelabs/open-operator-evals](https://github.com/nottelabs/open-operator-evals)\n\n# Agent features\n\n## Structured output\n\nStructured output is a feature of the agent's run function that allows you to specify a Pydantic model as the `response_format` parameter. The agent will return data in the specified structure.\n\n```python\nfrom notte_sdk import NotteClient\nfrom pydantic import BaseModel\n\nclass HackerNewsPost(BaseModel):\n    title: str\n    url: str\n    points: int\n    author: str\n    comments_count: int\n\nclass TopPosts(BaseModel):\n    posts: list[HackerNewsPost]\n\nclient = NotteClient()\nwith client.Session(open_viewer=True, browser_type=\"chrome\") as session:\n    agent = client.Agent(session=session, reasoning_model='gemini/gemini-2.5-flash', max_steps=15)\n    response = agent.run(\n        task=\"Go to Hacker News (news.ycombinator.com) and extract the top 5 posts with their titles, URLs, points, authors, and comment counts.\",\n        response_format=TopPosts,\n    )\nprint(response.answer)\n```\n\n## Agent vault\nVaults are tools you can attach to your Agent instance to securely store and manage credentials. The agent automatically uses these credentials when needed.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\n\nwith client.Vault() as vault, client.Session(open_viewer=True) as session:\n    vault.add_credentials(\n        url=\"https://x.com\",\n        username=\"your-email\",\n        password=\"your-password\",\n    )\n    agent = client.Agent(session=session, vault=vault, max_steps=10)\n    response = agent.run(\n      task=\"go to twitter; login and go to my messages\",\n    )\nprint(response.answer)\n```\n\n## Agent persona\n\nPersonas are tools you can attach to your Agent instance to provide digital identities with unique email addresses, phone numbers, and automated 2FA handling.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\n\nwith client.Persona(create_phone_number=False) as persona:\n    with client.Session(browser_type=\"chrome\", open_viewer=True) as session:\n        agent = client.Agent(session=session, persona=persona, max_steps=15)\n        response = agent.run(\n            task=\"Open the Google form and RSVP yes with your name\",\n            url=\"https://forms.google.com/your-form-url\",\n        )\nprint(response.answer)\n```\n\n# Session features\n\n## Stealth\n\nStealth features include automatic CAPTCHA solving and proxy configuration to enhance automation reliability and anonymity.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\n\n# Built-in proxies with CAPTCHA solving\nwith client.Session(\n    solve_captchas=True,\n    proxies=True,  # US-based proxy\n    browser_type=\"chrome\",\n    open_viewer=True\n) as session:\n    agent = client.Agent(session=session, max_steps=5)\n    response = agent.run(\n        task=\"Try to solve the CAPTCHA using internal tools\",\n        url=\"https://www.google.com/recaptcha/api2/demo\"\n    )\n\n# Custom proxy configuration\nproxy_settings = ExternalProxy(\n    server=\"http://your-proxy-server:port\",\n    username=\"your-username\",\n    password=\"your-password\",\n)\n\nwith client.Session(proxies=[proxy_settings]) as session:\n    agent = client.Agent(session=session, max_steps=5)\n    response = agent.run(task=\"Navigate to a website\")\n```\n\n## File download / upload\n\nFile Storage allows you to upload files to a session and download files that agents retrieve during their work. Files are session-scoped and persist beyond the session lifecycle.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\nstorage = client.FileStorage()\n\n# Upload files before agent execution\nstorage.upload(\"/path/to/document.pdf\")\n\n# Create session with storage attached\nwith client.Session(storage=storage) as session:\n    agent = client.Agent(session=session, max_steps=5)\n    response = agent.run(\n        task=\"Upload the PDF document to the website and download the cat picture\",\n        url=\"https://example.com/upload\"\n    )\n\n# Download files that the agent downloaded\ndownloaded_files = storage.list(type=\"downloads\")\nfor file_name in downloaded_files:\n    storage.download(file_name=file_name, local_dir=\"./results\")\n```\n\n## Cookies / Auth Sessions\n\nCookies provide a flexible way to authenticate your sessions. While we recommend using the secure vault for credential management, cookies offer an alternative approach for certain use cases.\n\n```python\nfrom notte_sdk import NotteClient\nimport json\n\nclient = NotteClient()\n\n# Upload cookies for authentication\ncookies = [\n    {\n        \"name\": \"sb-db-auth-token\",\n        \"value\": \"base64-cookie-value\",\n        \"domain\": \"github.com\",\n        \"path\": \"/\",\n        \"expires\": 9778363203.913704,\n        \"httpOnly\": False,\n        \"secure\": False,\n        \"sameSite\": \"Lax\"\n    }\n]\n\nwith client.Session() as session:\n    session.set_cookies(cookies=cookies)  # or cookie_file=\"path/to/cookies.json\"\n    \n    agent = client.Agent(session=session, max_steps=5)\n    response = agent.run(\n        task=\"go to nottelabs/notte get repo info\",\n    )\n    \n    # Get cookies from the session\n    cookies_resp = session.get_cookies()\n    with open(\"cookies.json\", \"w\") as f:\n        json.dump(cookies_resp, f)\n```\n\n## CDP Browser compatibility\n\nYou can plug in any browser session provider you want and use our agent on top. Use external headless browser providers via CDP to benefit from Notte's agentic capabilities with any CDP-compatible browser.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\ncdp_url = \"wss://your-external-cdp-url\"\n\nwith client.Session(cdp_url=cdp_url) as session:\n    agent = client.Agent(session=session)\n    response = agent.run(task=\"extract pricing plans from https://www.notte.cc/\")\n```\n\n# Hybrid workflows\n\nNotte's close compatibility with Playwright allows you to mix web automation primitives with agents for specific parts that require reasoning and adaptability. This hybrid approach cuts LLM costs and is much faster by using scripting for deterministic parts and agents only when needed.\n\n```python\nfrom notte_sdk import NotteClient\n\nclient = NotteClient()\n\nwith client.Session(open_viewer=True) as session:\n    # Start with a deterministic navigation\n    session.execute(type=\"goto\", url=\"https://duckduckgo.com/\")\n    session.execute(type=\"fill\", selector=\"internal:role=combobox[name=\\\"Search with DuckDuckGo\\\"i]\", value=\"nottelabs\")\n    agent = client.Agent(session=session, max_steps=3)\n    # Use an agent to reason about the next step\n    agent.run(task=\"Open nottelabs github repository\")\n    # Use a scraping endpoint to extract data\n    data = session.scrape(instructions=\"Extract number of stars\")\n```\n\n# Agent fallback\n\nWorkflows are a powerful way to combine scripting and agents to reduce costs and improve reliability. However, deterministic parts of the workflow can still fail. To gracefully handle these failures with agents, you can use the `AgentFallback` class: \n\n```python\nimport notte\n\nwith notte.Session() as session:\n    _ = session.execute(type=\"goto\", url=\"https://shop.notte.cc/\")\n    _ = session.observe()\n\n    with notte.AgentFallback(session, \"Go to cart\"):\n        # Force execution failure -\u003e trigger an agent fallback to gracefully fix the issue\n        res = session.execute(type=\"click\", id=\"INVALID_ACTION_ID\")\n```\n\n# Scraping\n\nFor fast data extraction, we provide a dedicated scraping endpoint that automatically creates and manages sessions. You can pass custom instructions for structured outputs and enable stealth mode.\n\n```python\nfrom notte_sdk import NotteClient\nfrom pydantic import BaseModel\n\nclient = NotteClient()\n\n# Simple scraping\nresponse = client.scrape(\n    url=\"https://notte.cc\",\n    scrape_links=True,\n    only_main_content=True\n)\n\n# Structured scraping with custom instructions\nclass Article(BaseModel):\n    title: str\n    content: str\n    date: str\n\nresponse = client.scrape(\n    url=\"https://example.com/blog\",\n    response_format=Article,\n    instructions=\"Extract only the title, date and content of the articles\"\n)\n```\n\nOr directly with cURL\n```bash\ncurl -X POST 'https://api.notte.cc/scrape' \\\n  -H 'Authorization: Bearer \u003cNOTTE-API-KEY\u003e' \\\n  -H 'Content-Type: application/json' \\\n  -d '{\n    \"url\": \"https://notte.cc\",\n    \"only_main_content\": false,\n  }'\n```\n\n\n**Search:** We've built a cool demo of an LLM leveraging the scraping endpoint in an MCP server to make real-time search in an LLM chatbot - works like a charm! Available here: [https://search.notte.cc/](https://search.notte.cc/)\n\n# License\n\nThis project is licensed under the Server Side Public License v1.\nSee the [LICENSE](LICENSE) file for details.\n\n# Citation\n\nIf you use notte in your research or project, please cite:\n\n```bibtex\n@software{notte2025,\n  author = {Pinto, Andrea and Giordano, Lucas and {nottelabs-team}},\n  title = {Notte: Software suite for internet-native agentic systems},\n  url = {https://github.com/nottelabs/notte},\n  year = {2025},\n  publisher = {GitHub},\n  license = {SSPL-1.0}\n  version = {1.4.4},\n}\n```\n\nCopyright © 2025 Notte Labs, Inc.\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnottelabs%2Fnotte","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fnottelabs%2Fnotte","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fnottelabs%2Fnotte/lists"}