KntroSoft is a software studio building ReqsAI — an AI copilot that turns live discovery meetings into structured, backlog-ready requirements.

KntroSoft is the organization behind ReqsAI, our flagship product. This org hosts ReqsAI's full stack (backend, frontend, infrastructure, marketing site) plus the academic work behind it. We're a small team — this is where the code, decisions, and architecture records live.

An analyst runs a live discovery session (voice or text), and an LLM listens in real time — surfacing draft user stories, edge cases, clarifying questions, and duplicate/update detection against the existing backlog as the conversation unfolds. Decisions are reviewed and accepted on the spot; the backlog stays structured, deduplicated, and exportable (including a direct push to Jira), with subscription billing via Stripe.

Try it: app.tamci.app

• Multi-tenant — schema-per-tenant isolation in a single Postgres instance.
• Modular monolith — Spring Modulith-enforced bounded contexts (iam, billing, workspace, discovery, gateway), each with an explicit public API surface; no reaching into another module's internals.
• Real-time — STOMP over WebSocket for live transcript streaming, AI suggestions, and session presence.
• Vector search — pgvector-backed similarity search for duplicate/related-story detection.
• Integrations — Jira Cloud (OAuth 2.0 + API token) for backlog push/import; Stripe for subscription billing.
• Production on AWS — ECS Fargate, RDS Postgres, CloudFront + S3, all provisioned via Terraform with GitHub Actions OIDC (no long-lived AWS keys).

We follow Gitflow across every repo: feature/* · bugfix/* · hotfix/* branches into develop, releases cut from develop into main. See each repo's .github/CONTRIBUTING.md for the full convention.

Questions, bug reports, or ideas? Open an issue on the relevant repo above — that's where we track everything.