Projects & Tools
A collection of tools, libraries, and open source projects I've built for software engineers and content creators. Some are free and open source -- others are commercial products I've built and continue to develop. My open source work is available on GitHub at github.com/ncosentino.
Tools I've Built
These are tools I've built and actively maintain as a software engineer and content creator.
BrandGhost
BrandGhost is the social media management platform I built to help creators and businesses show up everywhere, every day -- without the manual effort. It handles omni-channel scheduling, content recycling via Topic Streams, a unified comment feed, and an AI-powered content library.
Homebase
Homebase is a fully config-driven, SEO-optimized linktree-style landing page I built for developers who want full control over their personal landing page.
Fork it, edit one YAML file, and deploy automatically via GitHub Actions -- no runtime dependencies, no JavaScript required.
Includes built-in themes, Open Graph, JSON-LD, and llms.txt for AI answer engines.
.NET Libraries
Needlr
A source-generation-first dependency injection library for .NET with a fluent API. Supports AOT compilation, compile-time validation via Roslyn analyzers, decorator pattern, and keyed services. Available on NuGet as NexusLabs.Needlr.
MCP Servers
The Model Context Protocol (MCP) is an open standard that lets AI assistants like Claude and GitHub Copilot connect to external data sources and tools. These MCP servers are zero-dependency pre-built native binaries for Windows, macOS, and Linux -- no Node.js, Python, .NET, or Go runtime required.
Google Search Console MCP
A zero-dependency MCP server that connects AI assistants like Claude and GitHub Copilot to your Google Search Console data. Query search analytics, list sites, and inspect sitemaps -- all from within your AI workflow. Pre-built native binaries for Windows, macOS, and Linux.
Google Keyword Planner MCP
A zero-dependency MCP server that brings Google Ads Keyword Planner data into your AI assistant. Generate keyword ideas, get historical search volume, and forecast performance -- directly from Claude or GitHub Copilot. No runtime required.
Google PageSpeed Insights MCP
A zero-dependency MCP server that exposes Google PageSpeed Insights Core Web Vitals data to AI assistants. Analyze individual pages or batch-analyze multiple URLs and get actionable performance recommendations -- all from within your AI workflow.
Narnia
An MCP server and local web UI for browsing and searching your GitHub Copilot CLI session history. Exposes 9 MCP tools and indexes your local session store database, making it easy to retrieve past AI conversations and coding sessions.
Frequently Asked Questions
- What is the Model Context Protocol (MCP)?
- MCP (Model Context Protocol) is an open standard developed by Anthropic that lets AI assistants like Claude, GitHub Copilot, and others connect to external data sources and tools. An MCP server exposes a set of tools that the AI can invoke to retrieve data or perform actions -- for example, querying your Google Search Console data or analyzing a URL with PageSpeed Insights.
- What is Needlr?
- Needlr is a source-generation-first dependency injection library for .NET. Instead of using reflection at runtime, it generates the DI registration and resolution code at compile time using Roslyn source generators. This makes it fully AOT-compatible, faster at startup, and validated at compile time by built-in analyzers. It is available on NuGet as
NexusLabs.Needlr. - How do I connect Google Search Console to Claude or GitHub Copilot?
- Use the Google Search Console MCP server. It is a zero-dependency native binary -- download the release for your OS, configure your Google service account credentials, and add it to your MCP client configuration. No Node.js, Python, or .NET runtime is needed.
- Are these projects free and open source?
- Almost everything here is MIT-licensed and freely available on GitHub -- Needlr, all four MCP servers, and Homebase. The one exception is BrandGhost, which is a commercial product with its own pricing. Check brandghost.ai for details. Third-party APIs like Google Ads Keyword Planner have their own quota limits, but no limits are imposed by the open source projects themselves.
- What programming languages and runtimes are required?
- Needlr is a .NET NuGet package (C#, targeting .NET 8+). The four MCP servers (Google Search Console MCP, Google Keyword Planner MCP, Google PageSpeed Insights MCP, and Narnia) are distributed as pre-built native binaries for Windows, macOS, and Linux. No runtime -- no Node.js, no Python, no .NET, no Go -- is required to run them.