Policy and code,
seen together.
AI extracts claims from your policies and checks them against code, so you can investigate without waiting on an engineer.
Make the question specific.
A policy is a promise. See what supports it, what differs, and what still needs a person to check.
Explore a sample
Policy claim
“Delete activity records within 30 days.”
Retention policy, version 3
Code evidence
export const activityRetention = { retentionDays: 90, archiveBeforeDelete: false,}src/activity/retention.ts, pinned commit 7f3c8a1
The code specifies 90 days. A reviewer checks whether that meets the 30-day policy.
Planned
Your ChatGPT account.
Use the subscription you already have.
We’re working toward signing in with your ChatGPT account to use your existing subscription for Oculary’s AI and agents.
Oculary stores your uploaded documents and verification records. Hosted ChatGPT sign-in is not available yet.
From a claim
to a decision.
A shared path from a document to a reviewed finding, with the sources kept intact.
Find the claims
Upload a policy. AI automatically extracts claims that can be examined.
Follow the evidence
Investigate against an exact document version and repository commit. Ask an agent to look closer.
Make the review
Keep the reasoning, sources, and your decision together. Code evidence informs your judgment.
Keep the work moving.
Hand it to your team
Create and link a Jira task, or use Linear once its connector is set up. Keep ownership and the finding together.
Work toward a fix
Ask an agent to investigate and prepare a pull request. Review the change, then verify it against the requirement.
A closed task is progress. A fresh review establishes what the change actually supports.