{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"github_mcp-hangwin-mcp-chrome","slug":"mcp-hangwin-mcp-chrome","name":"mcp-chrome","type":"mcp","url":"https://github.com/hangwin/mcp-chrome","page_url":"https://unfragile.ai/mcp-hangwin-mcp-chrome","categories":["mcp-servers"],"tags":[],"pricing":{"model":"open_source","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"github_mcp-hangwin-mcp-chrome__cap_0","uri":"capability://tool.use.integration.mcp.protocol.bridging.via.native.messaging","name":"mcp protocol bridging via native messaging","description":"Exposes Chrome browser capabilities to external AI clients (Claude, etc.) through a Fastify-based Node.js server (mcp-chrome-bridge) running on port 12306 that implements the Model Context Protocol. Uses bidirectional JSON-RPC over Chrome native messaging to communicate between the extension and Node.js process, with Server-Sent Events (SSE) for streaming responses and STDIO as an alternative transport mechanism for clients that don't support HTTP.","intents":["Connect Claude or other MCP-compatible AI assistants to my Chrome browser for automation","Expose browser state and capabilities as tools to external AI agents","Stream real-time browser events and responses back to AI clients"],"best_for":["AI agent developers building Claude-integrated browser automation workflows","Teams deploying AI assistants that need persistent browser session access","Developers migrating from REST APIs to MCP-based tool orchestration"],"limitations":["Node.js server must run continuously on port 12306; no built-in clustering or load balancing","Native messaging adds ~50-100ms latency per round-trip vs direct extension APIs","STDIO transport requires manual process management; HTTP/SSE is recommended for production"],"requires":["Chrome browser with extension installed","Node.js 16+ for mcp-chrome-bridge","MCP-compatible AI client (Claude, etc.)","Port 12306 available on localhost"],"input_types":["JSON-RPC method calls","Tool invocation requests with parameters","MCP protocol messages"],"output_types":["JSON-RPC responses","Server-Sent Events (SSE) streams","Tool execution results"],"categories":["tool-use-integration","mcp-server"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_1","uri":"capability://automation.workflow.browser.interaction.recording.and.replay","name":"browser interaction recording and replay","description":"Captures user interactions (clicks, typing, navigation) in real-time and stores them as executable workflows in IndexedDB, enabling playback and modification through a visual workflow builder. Uses a transaction-based system to batch DOM mutations and event captures, with a flow data model that represents sequences of actions as nodes in a directed graph that can be executed, edited, and scheduled.","intents":["Record my browser interactions and replay them as automated workflows","Create reusable automation scripts without writing code","Schedule recorded workflows to run at specific times or on triggers"],"best_for":["Non-technical users automating repetitive browser tasks","QA teams creating regression test scripts visually","Business process automation teams building RPA workflows"],"limitations":["Recording captures DOM state at interaction time; dynamic content loaded after recording may not replay correctly","Complex JavaScript-driven interactions (drag-and-drop, custom gestures) may not record accurately","Workflows stored in IndexedDB are browser-local; no built-in cloud sync or cross-device sharing","Replay timing is deterministic but may fail if page structure changes between recording and playback"],"requires":["Chrome extension installed and running","IndexedDB support in Chrome (enabled by default)","Sufficient disk space for workflow storage (typically <1MB per workflow)"],"input_types":["User interactions (mouse events, keyboard input, navigation)","DOM state snapshots","Page screenshots"],"output_types":["Workflow JSON (flow data model)","Execution logs with timestamps","Screenshots from replay execution"],"categories":["automation-workflow","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_10","uri":"capability://data.processing.analysis.network.monitoring.and.request.interception","name":"network monitoring and request interception","description":"Captures and analyzes network requests made by the page, enabling workflows to wait for specific API calls, extract data from responses, or modify requests. Uses Chrome DevTools Protocol (CDP) to intercept network traffic, stores request/response metadata in the workflow context, and provides tools for conditional logic based on network events.","intents":["Wait for a specific API call to complete before proceeding with automation","Extract data from API responses for use in subsequent workflow steps","Verify that expected network requests are being made during automation"],"best_for":["QA teams testing API integrations through the UI","Developers debugging network issues in automation workflows","Teams building data extraction workflows that depend on API responses"],"limitations":["Network interception only works for requests made by the page; cannot intercept requests from extensions or other sources","Storing large response bodies in workflow context may consume significant memory","Request modification is limited to headers and body; cannot intercept HTTPS traffic without certificate installation","Network monitoring adds overhead; may slow down page load times on high-traffic pages"],"requires":["Chrome DevTools Protocol (CDP) support","Chrome extension with network monitoring permissions","Sufficient memory for storing request/response data"],"input_types":["Network request URLs","Request/response patterns","API response data"],"output_types":["Request/response logs","Extracted data from responses","Network timing information"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_11","uri":"capability://data.processing.analysis.offscreen.document.compute.for.ai.inference.and.media.encoding","name":"offscreen document compute for ai inference and media encoding","description":"Delegates compute-intensive operations (transformer model inference, GIF encoding, image processing) to an offscreen document that runs in a separate execution context, preventing blocking of the main UI thread. Uses Web Workers or offscreen document APIs to parallelize computation, with message passing to communicate results back to the main extension.","intents":["Run AI inference (embeddings, vision) without blocking the UI","Encode screenshots as GIFs for workflow playback without performance impact","Process large images or documents in the background"],"best_for":["Extension developers needing non-blocking compute","Teams running heavy ML models in the browser","Developers building responsive automation UIs"],"limitations":["Offscreen documents have limited API access; cannot access DOM or some Chrome APIs","Message passing between main and offscreen context adds latency (~5-10ms per message)","Memory is not shared; large data structures must be serialized/deserialized","Offscreen documents are terminated if the extension is disabled; no persistence"],"requires":["Chrome 109+ for offscreen document API","Manifest v3 extension","Sufficient system memory for parallel execution"],"input_types":["Screenshots (for encoding)","Text (for embedding)","Images (for processing)"],"output_types":["Encoded GIFs","Vector embeddings","Processed images"],"categories":["data-processing-analysis","image-visual"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_12","uri":"capability://automation.workflow.cli.interface.for.headless.workflow.execution","name":"cli interface for headless workflow execution","description":"Provides a command-line interface for executing recorded workflows in headless mode, enabling integration with CI/CD pipelines and server-side automation. Wraps the Node.js server with CLI commands for workflow execution, result reporting, and error handling, with support for parameterized workflows and output formatting.","intents":["Run automation workflows from CI/CD pipelines","Execute workflows on a server without a GUI","Integrate browser automation into command-line tools and scripts"],"best_for":["DevOps teams integrating automation into CI/CD","Developers building command-line tools for browser automation","Teams running scheduled automation on servers"],"limitations":["Headless execution requires a display server (Xvfb on Linux) or headless Chrome; not all pages render correctly in headless mode","CLI interface is limited to simple parameter passing; complex workflows may require code","Error messages from headless execution may be less informative than interactive debugging","No built-in support for parallel execution; workflows run sequentially"],"requires":["Node.js 16+","mcp-chrome-bridge running","Chrome or Chromium browser","Xvfb or similar display server (for Linux headless execution)"],"input_types":["Workflow ID","Parameters (JSON)","Output format (JSON, CSV, etc.)"],"output_types":["Execution results (JSON, CSV, etc.)","Exit codes","Logs"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_13","uri":"capability://automation.workflow.multi.tab.and.multi.window.coordination","name":"multi-tab and multi-window coordination","description":"Enables automation workflows to coordinate actions across multiple browser tabs and windows, with shared state management and cross-tab messaging. Uses Chrome extension message passing to synchronize state between tabs, enabling workflows that require interaction with multiple pages simultaneously or sequentially.","intents":["Automate workflows that require switching between multiple tabs","Coordinate actions across multiple browser windows","Share data between automation steps running in different tabs"],"best_for":["Teams automating complex multi-page workflows","Developers building workflows that require tab switching","QA teams testing multi-window interactions"],"limitations":["Cross-tab messaging adds latency; workflows may be slower than single-tab automation","State synchronization can be complex; race conditions may occur if multiple tabs modify shared state","Closed tabs lose their state; no automatic recovery for failed tab operations","Some pages use window.opener restrictions that prevent cross-window communication"],"requires":["Chrome extension with background service worker","Message passing infrastructure for cross-tab communication","Careful state management to avoid race conditions"],"input_types":["Tab IDs","Window IDs","Shared state data","Cross-tab messages"],"output_types":["Execution results from multiple tabs","Shared state snapshots","Coordination logs"],"categories":["automation-workflow","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_2","uri":"capability://image.visual.vision.based.browser.control.via.computertool","name":"vision-based browser control via computertool","description":"Enables AI agents to control the browser using visual perception by capturing screenshots, analyzing page layout, and executing actions (click, type, scroll) based on visual coordinates rather than DOM selectors. Implements a ComputerTool base class that accepts screenshot input, performs vision-based reasoning, and translates visual instructions into precise browser actions, supporting multi-step visual workflows.","intents":["Let Claude see and interact with my browser visually without needing DOM selectors","Automate complex UIs that use dynamic or obfuscated HTML","Enable AI agents to handle visual elements like images, charts, and custom components"],"best_for":["AI agents automating legacy or third-party web applications with unstable DOM","Teams building visual RPA workflows that don't require DOM-level precision","Developers integrating vision-language models (Claude's vision) with browser automation"],"limitations":["Vision-based control is slower than DOM-based interaction (requires screenshot + inference per action)","Coordinate-based clicking is fragile if page layout changes between screenshot and execution","Vision model must be able to interpret the visual content; fails on heavily obfuscated or non-standard UIs","No built-in OCR for text extraction; relies on vision model's text understanding"],"requires":["Vision-capable AI model (Claude 3.5+, GPT-4V, etc.)","Screenshot capability in browser (native Chrome API)","Sufficient API quota for vision inference"],"input_types":["Screenshots (PNG/JPEG)","Natural language instructions","Visual coordinates"],"output_types":["Click/type/scroll actions","Visual analysis results","Execution status with visual feedback"],"categories":["image-visual","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_3","uri":"capability://search.retrieval.semantic.similarity.search.with.onnx.based.embeddings","name":"semantic similarity search with onnx-based embeddings","description":"Provides vector-based semantic search over page content using transformer models (ONNX Runtime) running locally in the browser's offscreen document. Embeds page text into vector space using a pre-loaded model, stores vectors in an HNSW (Hierarchical Navigable Small World) index, and enables fast approximate nearest-neighbor search for finding relevant content without keyword matching.","intents":["Search for content on a page using semantic meaning, not just keywords","Find similar text passages across multiple pages or documents","Enable AI agents to locate relevant information without DOM selectors"],"best_for":["AI agents analyzing large documents or multi-page websites semantically","Teams building semantic search features without external vector databases","Developers needing privacy-preserving search (all inference local, no API calls)"],"limitations":["ONNX model inference adds 100-500ms latency per embedding depending on text length","HNSW index is in-memory; no persistence across browser sessions without manual export","Model size (typically 50-200MB) must fit in browser memory; larger models may cause slowdowns","Semantic search quality depends on model training; domain-specific content may have poor results"],"requires":["ONNX Runtime JavaScript library","Pre-trained transformer model (e.g., all-MiniLM-L6-v2)","Sufficient browser memory for model + index (typically 200-500MB)"],"input_types":["Text content from page","Query strings (natural language)","Page URLs or DOM elements"],"output_types":["Vector embeddings (float arrays)","Ranked search results with similarity scores","HNSW index data"],"categories":["search-retrieval","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_4","uri":"capability://planning.reasoning.real.time.agent.chat.with.streaming.tool.execution","name":"real-time agent chat with streaming tool execution","description":"Implements a streaming conversation interface where AI agents (Claude, etc.) can invoke browser tools in real-time and receive results within the same conversation flow. Uses a message processing pipeline that routes tool calls to the appropriate browser automation tools, captures results, and streams them back to the agent for multi-turn reasoning without waiting for full workflow completion.","intents":["Have a streaming conversation with Claude where it controls my browser in real-time","Let the AI agent see results of each action and adapt its next steps accordingly","Build multi-step workflows where the agent reasons about intermediate results"],"best_for":["AI agent developers building interactive automation workflows","Teams using Claude for complex, multi-step browser tasks requiring reasoning","Developers building chat interfaces for browser automation"],"limitations":["Streaming adds complexity to error handling; tool failures mid-stream may require conversation recovery","Agent reasoning latency compounds with tool execution time; complex workflows may take minutes","No built-in retry logic for failed tool calls; agent must explicitly request retry","Context window limits may require summarization for long-running conversations"],"requires":["MCP-compatible AI client with streaming support","mcp-chrome-bridge running and accessible","API key for AI model (Claude, etc.)"],"input_types":["Natural language instructions","Tool invocation requests","Conversation history"],"output_types":["Streaming text responses","Tool execution results","Conversation timeline with actions"],"categories":["planning-reasoning","tool-use-integration"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_5","uri":"capability://automation.workflow.visual.web.editor.with.shadow.dom.isolation","name":"visual web editor with shadow dom isolation","description":"Provides an in-browser visual editor overlay (built with React/Vue) that allows editing page content without interfering with the original page functionality. Uses shadow DOM to isolate the editor UI from page styles, implements a transaction-based batch apply system to commit edits atomically, and supports undo/redo through a state management system that tracks all mutations.","intents":["Edit page content visually without breaking page functionality","Test content changes before committing them to the page","Build visual editing workflows that can be recorded and replayed"],"best_for":["Content creators editing web pages visually","QA teams testing page modifications without code","Developers building no-code page customization tools"],"limitations":["Shadow DOM isolation prevents some CSS inheritance; custom styles may not apply correctly","Batch apply system requires careful transaction management; concurrent edits may conflict","Undo/redo is limited to the editor session; no persistence across page reloads without manual save","Complex JavaScript-driven pages may not respond correctly to DOM mutations from the editor"],"requires":["Chrome extension with visual editor entrypoint","React or Vue for UI rendering","Shadow DOM support (all modern browsers)"],"input_types":["DOM elements to edit","Text content","CSS styles"],"output_types":["Modified DOM","Edit transaction logs","Undo/redo state"],"categories":["automation-workflow","image-visual"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_6","uri":"capability://planning.reasoning.workflow.builder.with.node.based.flow.editor","name":"workflow builder with node-based flow editor","description":"Provides a visual node-based interface for constructing automation workflows where each node represents a browser action or decision point. Implements a flow data model that stores workflows as directed graphs with layout algorithms for automatic node positioning, supports conditional branching and loops, and integrates with the recording system to auto-generate nodes from captured interactions.","intents":["Build complex automation workflows visually without writing code","Create conditional logic (if/then) in automation workflows","Convert recorded interactions into editable, reusable workflows"],"best_for":["Non-technical users building RPA workflows","Business analysts designing process automation","Teams migrating from linear scripts to graph-based workflows"],"limitations":["Node layout algorithm may produce suboptimal layouts for very large workflows (100+ nodes)","Conditional logic is limited to simple branching; complex decision trees require nested nodes","No built-in version control; workflows are stored in IndexedDB without git-like history","Workflow execution is sequential; no built-in parallelization or async patterns"],"requires":["Chrome extension with workflow builder UI","React or similar framework for node editor","IndexedDB for workflow storage"],"input_types":["Recorded interactions","Manual node creation","Workflow templates"],"output_types":["Workflow JSON (flow data model)","Executable workflow definition","Visual diagram"],"categories":["planning-reasoning","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_7","uri":"capability://tool.use.integration.content.script.injection.and.dom.manipulation","name":"content script injection and dom manipulation","description":"Injects content scripts into web pages to capture user interactions, monitor DOM changes, and execute browser automation commands at the page level. Uses a message passing architecture to communicate between content scripts and the background service worker, enabling real-time event capture (clicks, typing, navigation) and DOM mutations without blocking page execution.","intents":["Capture user interactions on any web page for recording","Execute automation commands (click, type, scroll) on the current page","Monitor page changes and trigger workflows based on DOM mutations"],"best_for":["Browser extension developers building automation tools","Teams needing real-time page interaction capture","Developers building event-driven automation workflows"],"limitations":["Content scripts cannot access cross-origin iframes; automation is limited to same-origin content","Message passing between content script and background worker adds ~10-20ms latency per command","DOM mutation monitoring can be expensive on pages with frequent updates; may cause performance degradation","Some pages use Content Security Policy (CSP) that blocks script injection; requires manifest configuration"],"requires":["Chrome extension manifest with content_scripts and host_permissions","Target website must not have restrictive CSP","Background service worker for message routing"],"input_types":["DOM events (click, input, change)","Page navigation events","Automation commands (click, type, scroll)"],"output_types":["Event logs with timestamps","DOM state snapshots","Command execution results"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_8","uri":"capability://memory.knowledge.project.and.session.management.with.sqlite.persistence","name":"project and session management with sqlite persistence","description":"Manages multiple automation projects and conversation sessions using a SQLite database with Drizzle ORM, enabling users to organize workflows, save conversation history, and switch between different automation contexts. Stores project metadata, workflow definitions, execution logs, and conversation transcripts in a structured relational schema accessible through the Node.js server.","intents":["Organize multiple automation workflows into projects","Save and resume conversations with AI agents across sessions","Track execution history and logs for debugging and auditing"],"best_for":["Teams managing multiple automation projects","Developers building persistent AI agent applications","Organizations requiring audit trails for automated processes"],"limitations":["SQLite is single-writer; concurrent access from multiple processes may cause lock contention","Database is local to the machine; no built-in cloud sync or multi-device access","Drizzle ORM adds abstraction overhead; complex queries may be slower than raw SQL","No built-in backup or disaster recovery; data loss if database file is corrupted"],"requires":["Node.js 16+ with SQLite support","Drizzle ORM installed","Disk space for database (typically <100MB per 1000 workflows)"],"input_types":["Project metadata (name, description)","Workflow definitions","Conversation messages","Execution logs"],"output_types":["Project list","Workflow history","Conversation transcripts","Execution reports"],"categories":["memory-knowledge","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"github_mcp-hangwin-mcp-chrome__cap_9","uri":"capability://automation.workflow.trigger.system.and.workflow.scheduling","name":"trigger system and workflow scheduling","description":"Enables scheduling and triggering of recorded workflows based on time intervals, events, or external conditions. Implements a trigger registry that maps conditions (cron expressions, webhook events, page changes) to workflow execution, with a scheduler that manages timing and ensures reliable execution even if the browser is closed.","intents":["Schedule workflows to run at specific times (e.g., daily reports)","Trigger workflows when specific events occur (page change, webhook)","Automate repetitive tasks without manual intervention"],"best_for":["Teams automating scheduled tasks (daily reports, data collection)","Developers building event-driven automation","Organizations needing reliable workflow scheduling"],"limitations":["Scheduling requires the Node.js server to be running continuously; no built-in high availability","Cron-based scheduling is limited to time-based triggers; complex event logic requires custom code","Workflow execution is not guaranteed if the browser is closed or network is unavailable","No built-in retry logic for failed executions; requires manual intervention or custom handlers"],"requires":["mcp-chrome-bridge running continuously","Cron expression library (e.g., node-cron)","Webhook endpoint for event-based triggers (optional)"],"input_types":["Cron expressions","Webhook payloads","Event conditions","Workflow IDs"],"output_types":["Execution logs","Trigger status","Workflow results"],"categories":["automation-workflow","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":48,"verified":false,"data_access_risk":"high","permissions":["Chrome browser with extension installed","Node.js 16+ for mcp-chrome-bridge","MCP-compatible AI client (Claude, etc.)","Port 12306 available on localhost","Chrome extension installed and running","IndexedDB support in Chrome (enabled by default)","Sufficient disk space for workflow storage (typically <1MB per workflow)","Chrome DevTools Protocol (CDP) support","Chrome extension with network monitoring permissions","Sufficient memory for storing request/response data"],"failure_modes":["Node.js server must run continuously on port 12306; no built-in clustering or load balancing","Native messaging adds ~50-100ms latency per round-trip vs direct extension APIs","STDIO transport requires manual process management; HTTP/SSE is recommended for production","Recording captures DOM state at interaction time; dynamic content loaded after recording may not replay correctly","Complex JavaScript-driven interactions (drag-and-drop, custom gestures) may not record accurately","Workflows stored in IndexedDB are browser-local; no built-in cloud sync or cross-device sharing","Replay timing is deterministic but may fail if page structure changes between recording and playback","Network interception only works for requests made by the page; cannot intercept requests from extensions or other sources","Storing large response bodies in workflow context may consume significant memory","Request modification is limited to headers and body; cannot intercept HTTPS traffic without certificate installation","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.6703665631061988,"quality":0.5,"ecosystem":0.39999999999999997,"match_graph":0.25,"freshness":0.6,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.15,"match_graph":0.23,"freshness":0.12}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-05-24T12:16:22.065Z","last_scraped_at":"2026-05-03T14:23:31.492Z","last_commit":"2026-01-06T13:05:39Z"},"community":{"stars":11371,"forks":1014,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=mcp-hangwin-mcp-chrome","compare_url":"https://unfragile.ai/compare?artifact=mcp-hangwin-mcp-chrome"}},"signature":"jHpI1PYN1M2+Cv/wSxJ1JK1nlvEOhZeznC8AXPqzxCwzPCxtFb3vSmvnQQrZU20xIrtKWAjzeMjp4berO286Ag==","signedAt":"2026-06-20T08:26:00.188Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/mcp-hangwin-mcp-chrome","artifact":"https://unfragile.ai/mcp-hangwin-mcp-chrome","verify":"https://unfragile.ai/api/v1/verify?slug=mcp-hangwin-mcp-chrome","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}