An MCP server providing tools to control web browsers using the Amazon Nova Act SDK
Listed on its own since 2026-09-25. Until then it was part of iflow-mcp-madtank-nova-act-mcp-server, where its earlier checks remain.
Wellknown found it in public sources; nobody has proven control of it yet. Claiming takes one click if the repository is under your GitHub account, or a small file on your domain otherwise. Verified owners get the badge, 15-minute checks, status alerts, edits that outrank crawled data, and a ranking boost.
Agents can do it too: POST https://wellknown.network/api/v1/claims with {"agent":"nova-act-mcp-server","method":"well_known_file"} — machine-readable steps at claim.json, guide at /docs/claim.
Everything here was measured by our prober or read from a registry. Nothing is self-reported.
Attributed to the source that supplied each field. Treated as claims, not facts.
# nova-act-mcp-server [](https://pypi.org/project/nova-act-mcp-server/) **nova‑act‑mcp‑server** is a zero‑install [Model Context Protocol](https://modelcontextprotocol.io/) (MCP) server that exposes [Amazon Nova Act](https://nova.amazon.com/act) browser‑automation tools for AI agents. ## What's New in v3.2.0 - **File-Based Screenshots**: Screenshots now save to files instead of inline base64, eliminating MCP message size limits - **Optimized Image Quality**: Reduced screenshot quality (60/30) for smaller file sizes (~50KB vs 500KB+) - **Enhanced Error Handling**: Improved exception handling for browser operations - **Better Testing**: Comprehensive test coverage for screenshot functionality - **MCP Inspector Integration**: New testing script for easy validation ## Quick Start (uvx) ### Step 1: Get a Nova Act API Key Obtain your API key from [Nova Act](https://nova.amazon.com/act). ### Step 2: Add to MCP Client Configuration ```jsonc { "mcpServers": { "nova_act_browser_tools": { "command": "uvx", "args": ["nova-act-mcp-server"], "env": { "NOVA_ACT_API_KEY": "YOUR_NOVA_ACT_API_KEY_HERE" } } } } ``` ### Step 3: Start Using Browser Tools AI agents can now call tools like `start_session`, `execute_instruction`, etc. through any MCP-compatible client such as Claude Desktop or VS Code. ## Core Tools Overview | Tool | Description | Key Parameters | |------|-------------|----------------| | `start_session` | Starts a new browser session | `url`, `headless=True` | | `execute_instruction` | Runs a natural language instruction in the browser | `session_id`, `instruction` | | `inspect_browser` | Gets a screenshot and state of the current browser | `session_id`, `include_screenshot=True` | | `end_session` | Closes a browser session and cleans up resources | `session_id` | | `list_browser_sessions` | Lists all active browser sessions | None | ## Local Development & Testing ### Se…
Mapped onto the structured taxonomy from declared text and observed tool names. Confidence shown for derived entries.
Every source is kept verbatim. Field changes are logged as events.