Show HN: Cont3xt.dev – Universal Team Knowledge for AI Coding Tools

Hi HN! I am a huge user of AI coding tools, especially Claude Code and Cursor, but every session starts fresh. It doesn't remember yesterday's architectural decision or know your team's coding patterns.

I built Cont3xt.dev to give Claude (and other AI tools) persistent access to team context via MCP.

*How it works:*

1. Document your team's context once (architectural decisions, coding standards, patterns)

2. When Claude Code (or Claude.ai, Cursor, etc.) needs context, it fetches relevant info via MCP

3. Claude generates code that actually matches your team's approach

*Example:* You: "Create a user service" Without context: Claude might use different patterns than your existing services With Cont3xt: Claude knows you use repository pattern, knows your error handling approach, follows your naming conventions

*Features:*

- Rules library (prioritised - critical rules always included)

- Architectural Decision Records (why you chose X over Y)

- Smart filtering (only relevant context, respects token budgets)

- Analytics (track what's actually useful)

Looking for beta testers and lots of feedback. Free during beta.

Cont3xt.dev is what every threat intelligence platform should be: brutally efficient. As a security analyst, I've wasted countless hours juggling between VirusTotal, Shodan, and passive DNS tools. This platform finally brings everything together in one clean interface The relationship mapping is genuinely useful - not just visual fluff. I can trace an IP to domains, then to certificates and back to other IPs in seconds. The automated enrichment saves me 10-15 minutes per investigation My only complaint? I wish I'd found this tool years earlier. It's become as essential as my password manager. Simple, focused, and actually delivers on its promise without the enterprise bloat