network-mcp
Enables supervised web form assistance with ChatGPT, opening pages visibly, discovering opportunities, and requiring user approval before any fill actions.
README
network-mcp
network-mcp is a ChatGPT-driven, supervised opportunity-assistance MVP for Windows.
It combines:
- a thin Cloudflare Worker gateway
- a local visible Playwright browser worker
- a small MCP tool registry skeleton
- explicit approval before fill actions
- submit disabled by default
MVP flow
- User provides one URL in ChatGPT.
- The browser worker opens it visibly.
- The worker discovers likely opportunity links within a limited scope.
- ChatGPT presents options.
- The user selects a page or form.
- The worker opens it visibly and inspects fields.
- ChatGPT proposes values.
- The user approves.
- The worker fills approved fields only.
- Submit stays manual unless an explicit submit mode is enabled later.
Final submit is intentionally not auto-implemented in the MVP.
Policy defaults
NETWORK_MCP_MODE=local-assist
NETWORK_MCP_HEADLESS=false
NETWORK_MCP_REQUIRE_APPROVAL_FOR_FILL=true
NETWORK_MCP_REQUIRE_APPROVAL_FOR_SUBMIT=true
NETWORK_MCP_SUBMIT_DEFAULT=false
NETWORK_MCP_ENABLE_SUBMIT=false
NETWORK_MCP_MAX_SESSION_SECONDS=900
NETWORK_MCP_MAX_PAGE_VISITS=20
NETWORK_MCP_MAX_FORM_FILLS=1
NETWORK_MCP_MAX_FIELD_WRITES=80
NETWORK_MCP_ALLOWED_HOSTS=
NETWORK_MCP_DENIED_HOSTS=
NETWORK_MCP_PUBLIC_ORIGIN=
NETWORK_MCP_WORKER_PORT=8791
NETWORK_MCP_BROWSER_CHANNEL=msedge
NETWORK_MCP_EXTERNAL_VISIBLE_CHROME=false
Tool contract
Documented safe operations:
network.opennetwork.inspectnetwork.extract_formnetwork.proposenetwork.fill_after_approvalnetwork.review_before_submit
Safety defaults
- Browser is visible by default.
- Login and registration are user-driven.
- CAPTCHA, 2FA, and security challenge pages pause automation.
- Credentials are entered manually by the user.
- Form values are not logged by default.
- Browser storage, cookies, screenshots, traces, videos, and downloads stay in ignored runtime directories such as
var\browser\. - One active browser session is the default operating model.
Local commands
cd D:\PhpstormProjects\www\network-mcp
npm install
pwsh -NoProfile -ExecutionPolicy Bypass -File .\tool\dev-network.ps1 doctor
pwsh -NoProfile -ExecutionPolicy Bypass -File .\tool\dev-network.ps1 start
pwsh -NoProfile -ExecutionPolicy Bypass -File .\tool\dev-network.ps1 smoke-local
Documentation
Cloudflare
- Worker config:
cloudflare-worker/wrangler.jsonc - Public health check:
/healthzreturnsservice: "network-mcp" - Worker deploy remains a manual action through the supervisor or
wrangler
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.