quint-code

quint-code

quint-code

Structured reasoning for AI coding tools

Make better decisions. Remember why you made them.

$ curl -fsSL https://quint.codes/install.sh | bash -s -- -g
Get Started View on GitHub
Claude Code • Cursor • Gemini CLI • Codex CLI

You're deep in a codebase. You need to make an architectural decision.
Event choreography? Saga pattern? Outbox with CDC?

Your AI assistant gives you an answer. It's coherent. But:

"Why this approach?"

You won't remember the reasoning in 3 months. The decision lives in a chat thread you'll never find again.

"What alternatives?"

Were there alternatives, or did you anchor on the first idea? No record of what was considered and rejected.

"When does it expire?"

The tradeoffs shift as the system evolves. No one tracks when assumptions become invalid.

The ADI Cycle

Abduction → Deduction → Induction → Audit → Decision

AI generates options. You decide. Everything is documented.

1

Hypothesize

Generate 3-5 competing approaches

2

Verify

Check logical consistency

3

Test

Gather evidence: benchmarks, docs

4

Audit

WLNK analysis, check blind spots

5

Decide

Create Design Rationale Record

Before & After

Without quint-code

  • × "Let's use Saga pattern" (first idea)
  • × "Why did we build it this way?" (3 months later)
  • × "Is this assumption still valid?"
  • × Same debates in different meetings

With quint-code

  • ✓ 3 approaches analyzed, tradeoffs documented
  • ✓ Check DRR-001.md — full rationale preserved
  • ✓ Run /q-decay — evidence freshness tracked
  • ✓ Query knowledge base — past decisions searchable

See it in action

quint-code
$ /q1-hypothesize "How should we handle cross-service transactions?"
Generating hypotheses...
H1: Saga pattern with orchestrator
Central coordinator manages the workflow
H2: Event choreography
Services react to domain events, no coordinator
H3: Outbox pattern with CDC
Transactional outbox, Change Data Capture for publishing
Hypotheses saved to .quint/knowledge/L0/
Next: Run /q2-verify to check logical consistency

From the creator of quint-code

ivan zakutni

First Principles Engineering

Production AI systems. Software architecture decisions that don't age like milk. The same methodology behind quint-code — applied to real-world engineering problems.

Deep dives, not hot takes. For engineers who ship.