Show HN: Straion – dynamic AGENTS.md/Claude.md context for AI coding agents (straion.com)
Core idea: static AGENTS.md / CLAUDE.md files are useful, but they often become too broad for real tasks. Instead of injecting all guidance every time, Straion selects the relevant rules per task and keeps context narrow.
How it works today:
1. You define org/team/project rules centrally 2. A CLI skill calls our pipeline during agent workflows 3. Straion returns only the rules relevant to the current task 4. The agent gets focused context instead of a giant rules dump
This reduces context bloat and helps avoid “silent rule misses” in coding-agent outputs (Claude Code, Cursor, Copilot workflows).
We’re launching today on Product Hunt and opened a free tier for everyone: https://www.producthunt.com/products/straion
Happy to share details on architecture/tradeoffs and get roasted on where this breaks.
1 comment
[ 3.3 ms ] story [ 10.5 ms ] threadWe built Straion because AI-generated code is everywhere — but in reality, it rarely fits how companies actually build software.
The problem isn’t generating code anymore. It’s alignment. Every company has its own standards for security, privacy, architecture, design systems, and frameworks. Yet AI tools don’t automatically understand those rules. The result? Manual fixes, long review cycles, and wasted time.
We built Straion to change that.
Straion automatically extracts company-specific requirements from sources like wikis, contribution guidelines, and best practices — and translates them into instructions AI agents can actually follow. That way, generated code fits the organization from the start.
This means:
Less manual correction
Fewer review loops
Better security and compliance alignment
Faster, more cost-efficient delivery
Before building, we conducted 100+ interviews with software teams to truly understand their pain points. The result is a product that doesn’t just work technically — it solves a real, expensive problem.
Ultimately, we built Straion so developers can focus on what really matters again: building great software instead of fixing AI output.