← All projects

Comp AI (v2)

Comp AI re-architected. Cleaner data model, refined RBAC, structured AI integrations. Useful diff target vs v1.

Category
vertical-saas
Language
TypeScript
Runtime
Node · Postgres
Providers
Anthropic, OpenAI
Agent loop
n/a
Tool format
n/a
Memory
n/a
Caching
auto
Sandbox
none
Pedagogical
★★★☆☆
Flags
Concepts
provider-abstraction ·guardrails
Verified
paper-pass
Upstream

What it is

The v2 of Comp AI — a compliance-as-a-service vertical SaaS rebuilt with a cleaner data model, refined RBAC, and structured AI integrations. The most directly relevant codebase in the corpus to a GRC platform.

What’s worth studying

  • Tenant-scoped data model. RBAC is a first-class table, not bolted onto a generic auth library. The migrations show how to evolve a tenant model without breaking existing tenants — relevant when your AI-Act tier classification changes mid-product.
  • Structured AI integrations. AI-generated content (policies, risk assessments) goes through a typed schema before it reaches the customer. Hallucination is bounded by validation; auditing becomes trivial.
  • The diff against v1. What was renamed, what was removed, what got pluggable. Useful pattern-matching for any v2 you’ll do yourself.

Drill-down

The full per-doc analysis lives below — these are the original numbered analyses, rendered as styled HTML. Pick a section to study deeper.