Can I do my whole SOC 2 audit inside Claude or ChatGPT?
Mostly yes. Readiness and evidence collection run inside an AI agent on your computer: Claude Code, Codex (OpenAI's agent), Cursor, Gemini CLI or the Grok bot. You sign agreements and upload your evidence in your portal, and our audit team examines it. It ends in a signed SOC 2 report, not a dashboard.
What happens inside your AI
- Your AI checks your setup against our published checks and walks you through each fix.
- It gathers copies of your evidence into one folder on your computer.
- It runs read-only commands where it can, and you approve each one.
What happens outside it
- You sign agreements and upload your audit folder in your portal.
- Our audit team examines the evidence, and Y Assurance PLLC issues your report. Your AI never decides whether you pass.
Why not the ChatGPT or Claude chat app?
A chat app in your browser can't reach the files on your computer. This workflow needs an agent that can read them and run evidence-collection commands, with your approval at each step.
How to connect
Add one address to your agent, then sign in:
https://mcp.chiarohq.com/mcp
Setup steps for each tool are on our MCP page.
Questions
Does the AI audit me?
No. Your AI gathers and organizes the evidence. Our audit team examines it, and Y Assurance PLLC issues the report.
Do you get my passwords or keys?
No. Nothing is installed on your systems, and you approve every command your AI runs.
Where does my evidence go?
During readiness, your files stay on your computer and we keep a record of each one. For the audit, you upload copies to your evidence vault.