> ## Documentation Index > Fetch the complete documentation index at: https://gologin.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Gologin for AI Agents > Give your AI agents reliable web access — scraping, browser automation, and persistent sessions through Gologin Gologin gives AI agents three things they need to work with the web reliably: 1. **Read the web** — scrape any page, even JS-heavy and protected sites, through Web Unlocker 2. **Interact with the web** — open real browser sessions with clicks, typing, navigation, and persistent cookies through Cloud Browser 3. **Stay undetected** — anti-detect fingerprinting, residential proxies, and profile-based identity management These capabilities are available at multiple integration levels — pick what fits your stack. *** ## Integration options Connect Gologin to Claude Desktop, Cursor, or any MCP client. Manage profiles and sessions through natural language. Plug-and-play extensions for Claude Code. Install a skill and your agent gets web scraping and browser automation out of the box. Shell commands that AI agents call directly. Scrape, browse, interact — no code to write. Three CLIs for different workflows. Full programmatic control. Build custom agent tools with Puppeteer/Playwright connected to Gologin Cloud Browser. *** ## Which tool to choose | I want to... | Use | | --------------------------------------------- | ----------------------------------------------------------------- | | Manage profiles from Claude Desktop / Cursor | [MCP Server](/api-reference/ai-integrations/mcp-server) | | Give Claude Code web access with zero setup | [AI Skills](/api-reference/ai-integrations/ai-skills) | | Scrape pages, crawl sites, batch extract | [Web Access CLI](/cli-tools/web-access) | | Open a cloud browser, click, type, screenshot | [Agent Browser CLI](/cli-tools/agent-browser) | | Warm up local profiles, run login flows | [Local Agent Browser CLI](/cli-tools/local-agent-browser) | | Build custom tools with Puppeteer/Playwright | [Cloud Browser API](/api-reference/cloud-browser/getting-started) | | Call the scraping API from my own code | [Web Unlocker SDK](/scraping-api-sdk-and-cli) | *** ## How it all fits together ``` ┌─────────────────────────────────────────────────┐ │ AI Agent │ │ (Claude, Codex, custom agent) │ └──────┬──────────┬──────────┬────────────────────┘ │ │ │ MCP Server AI Skills CLI Tools / SDK │ │ │ └──────────┴──────────┘ │ ┌─────────────┴─────────────┐ │ │ Web Unlocker Cloud Browser (stateless scraping) (stateful sessions) │ │ └─────────────┬─────────────┘ │ Gologin Infrastructure (proxies, fingerprints, profiles) ``` * **Web Unlocker** handles stateless read/extraction — send a URL, get rendered HTML back * **Cloud Browser** handles stateful interaction — open a session, click, type, navigate, persist cookies * **Local Orbita** handles persistent local profiles — warmup, login flows, social/marketplace actions All three share the same Gologin infrastructure: residential proxies, anti-detect fingerprinting, and profile management. *** ## What people build with this ### Research and competitive intelligence An AI agent that monitors competitor pricing, product launches, and content changes across dozens of websites — daily, automatically. The agent scrapes competitor pages through Web Unlocker, extracts structured data, compares with yesterday's snapshot, and produces a daily digest. No browser sessions needed, just stateless extraction at scale. ```bash theme={null} # scrape 20 competitor pricing pages and get structured output gologin-web-access batch-scrape \ https://competitor1.com/pricing \ https://competitor2.com/pricing \ https://competitor3.com/pricing \ --format json --summary --output ./today-prices.json # detect what changed since last check gologin-web-access batch-change-track \ https://competitor1.com/pricing \ https://competitor2.com/features \ --format markdown --output ./changes.json ``` ### Lead generation and enrichment Scrape business directories, LinkedIn company pages, review sites — then feed the HTML into Claude to extract names, emails, phone numbers, company size, tech stack. Works as a tool inside an AI agent loop: the agent decides which pages to scrape, calls the tool, gets structured data back. ```bash theme={null} # search for leads gologin-web-access search "saas companies series A 2026" --limit 20 # scrape and extract structured data from each result gologin-web-access extract https://example.com/company/123 --schema ./lead-schema.json ``` ### Multi-account management Run 50+ social media accounts, ad accounts, or marketplace seller accounts — each with its own browser profile, fingerprint, proxy, and cookies. Local Agent Browser manages the profiles, warms them up with runbooks, and keeps sessions alive across agent calls. ```bash theme={null} # create profiles for each account gologin-local-agent-browser profile-create "FB Ads Account 1" --template ads gologin-local-agent-browser profile-create "FB Ads Account 2" --template ads # warm up profiles in batch gologin-local-agent-browser batch ./warmup-runbook.json \ --targets ./all-profiles.json --concurrency 5 ``` ### Automated form filling and data entry An AI agent that logs into internal dashboards, fills forms, submits reports, downloads exports — all through a real cloud browser session. The agent reads the page as a snapshot, finds the right fields by ref, types values, clicks submit, and verifies the result. ```bash theme={null} gologin-agent-browser open https://internal-dashboard.com/login --profile my_profile gologin-agent-browser snapshot gologin-agent-browser fill @e3 "username" gologin-agent-browser fill @e5 "password" gologin-agent-browser click @e7 # submit button gologin-agent-browser wait 2000 gologin-agent-browser snapshot # verify login success gologin-agent-browser screenshot ./logged-in.png ``` ### Content monitoring and alerting Track changes on government regulation pages, competitor blogs, job boards, or any page that matters to your business. Run daily via cron or a scheduled agent, diff against previous version, alert on meaningful changes. ```bash theme={null} # track changes on key pages gologin-web-access change-track https://regulations.gov/new-rules --format markdown gologin-web-access batch-change-track \ https://competitor.com/blog \ https://competitor.com/changelog \ --output ./changes.json ``` ### AI-powered web scraping pipelines Build an end-to-end pipeline: crawl a docs site, convert to markdown, feed into an LLM for summarization or Q\&A. Use the CLI for extraction, and your own code for the LLM part. ```bash theme={null} # crawl an entire docs site gologin-web-access crawl https://docs.example.com \ --format markdown --limit 100 --max-depth 3 --only-main-content \ --output ./docs-corpus.json # or map the site first, then selectively scrape gologin-web-access map https://docs.example.com --limit 200 gologin-web-access batch-scrape url1 url2 url3 --format text --output ./selected.json ``` ### Geo-testing and localization QA Check how your website looks from different countries. Open the same URL through profiles configured with proxies in different geos, take screenshots, compare. ```bash theme={null} # create geo-specific profiles gologin-local-agent-browser profile-create "US Test" --template geo --proxy-country us gologin-local-agent-browser profile-create "DE Test" --template geo --proxy-country de gologin-local-agent-browser profile-create "JP Test" --template geo --proxy-country jp # screenshot from each location gologin-agent-browser open https://yoursite.com --profile us_profile_id gologin-agent-browser screenshot ./us-view.png gologin-agent-browser close ``` *** ## Quick start examples ### Scrape a page (CLI) ```bash theme={null} npm install -g gologin-web-access export GOLOGIN_WEB_UNLOCKER_API_KEY="wu_..." gologin-web-access read https://example.com ``` ### Open a cloud browser (CLI) ```bash theme={null} npm install -g gologin-agent-browser-cli export GOLOGIN_TOKEN='your_token' gologin-agent-browser open https://example.com gologin-agent-browser snapshot gologin-agent-browser click @e3 gologin-agent-browser close ``` ### Install an AI Skill (Claude Code) ```bash theme={null} npx skills add GologinLabs/gologin-web-access-skill ``` Then just ask Claude Code to scrape a page or interact with a site — the skill handles the rest. ### Connect MCP (Claude Desktop) Add to `claude_desktop_config.json`: ```json theme={null} { "mcpServers": { "gologin-mcp": { "command": "npx", "args": ["gologin-mcp"], "env": { "API_TOKEN": "your-token" } } } } ``` Then ask Claude: "Create a new browser profile with a US proxy."