wolfgrey // Claude — Our AI Stack

The Foundation

Why Claude?

After testing every major AI model, we standardized on Claude for everything we build. Here's why.

Best for Business Writing

Claude produces the most natural, professional prose. No AI-speak, no weird phrasing. Output that sounds like a skilled human wrote it.

Honest About Limits

Claude tells you when it doesn't know something instead of making things up. Critical when you're making business decisions based on AI output.

Massive Context Window

200K tokens means Claude can read entire documents, codebases, or conversation histories. No artificial memory limits on complex projects.

Safety-First Design

Anthropic's constitutional AI approach means Claude follows instructions carefully without unpredictable behavior. Reliable enough for production workflows.

Developer-Friendly

Clean API, excellent documentation, and tools like Claude Code make it easy to build real applications. Not just a chatbot.

Aligned With Our Values

Anthropic's mission is building safe, beneficial AI. We want to work with companies thinking long-term about AI's impact on society.

Our Stack

The tools we use daily

From conversational AI to autonomous coding agents, here's what powers wolfgrey.

Claude (claude.ai)

The conversational interface for research, writing, and thinking

Claude.ai is our daily driver for anything that requires thinking through a problem, drafting content, or exploring ideas. We use it for:

Client research — Understanding industries, competitors, and market dynamics before engagements
Content creation — First drafts of guides, emails, proposals, and documentation
Strategy sessions — Thinking through implementation approaches and trade-offs
Learning — Getting up to speed on new tools, frameworks, and concepts quickly

The Pro plan gives us access to Claude's most capable models and priority access during high-demand periods.

Try Claude free →

Claude Code

Anthropic's agentic coding tool that lives in your terminal

Claude Code is a game-changer. It's an AI coding agent that can read your codebase, make changes across multiple files, run commands, and iterate on its own work. Unlike chat-based coding assistants, it actually does the work.

Full codebase awareness — It reads and understands your entire project, not just the file you're looking at
Multi-file edits — Makes coordinated changes across multiple files in a single operation
Terminal access — Runs tests, installs packages, commits code, and executes commands
Agentic loops — Can work autonomously, trying approaches, checking results, and iterating

We use Claude Code to build client automations, create internal tools, maintain this website, and prototype new ideas. It's like having a senior developer who never gets tired.

Learn about Claude Code →

Model Context Protocol (MCP)

The open standard that connects Claude to external tools and data

MCP is how we give Claude superpowers. It's an open protocol that lets Claude connect to databases, APIs, file systems, and other tools securely. Think of it as USB for AI.

Supabase integration — Claude can query and modify databases directly
Browser automation — Control Chrome to scrape data, fill forms, or test interfaces
File system access — Read and write files, manage projects, organize data
Custom tools — Build bespoke integrations for specific client needs

MCP is what makes our AI agents actually useful. Without it, Claude is just a chatbot. With it, Claude becomes a capable assistant that can take real actions.

Explore MCP →

Why we chose Claude