Submit architecture input
Start with an architecture brief, markdown design, or infrastructure context for the system you are evaluating.
Cloud architecture evaluator
Dry run your architecture before production does.
Submit architecture decisions for senior-style evaluation. dryrun challenges assumptions with targeted cross-verification questions, then generates ADR drafts from finalized design choices.
Cross-verification questions
Risk evidence checks
ADR output
$ dryrun evaluate architecture.md
Architecture Evaluation Report
Input: Multi-tenant SaaS on AWS
Context: 50k users, tenant isolation, async jobs, sane cost.
Cross-verify: where is noisy-neighbor isolation enforced?
ALB -> ECS services -> RDS Postgres. SQS for background jobs. Shared tables use tenant_id with query-level filtering.
Follow-up: what prevents long-running tenant reports from exhausting DB connections for interactive traffic?
Reliability
Needs work
Evidence
No workload isolation, no queue classing, and no stated retry or DLQ strategy.
Next question
Where does backpressure happen when reporting traffic spikes, and what SLO protects read latency?
ADR generated
ADR-004: Queue-based report offloading with tenant workload isolation, retry policy, and DLQ boundaries.
ALB
ECS
RDS
SQS
Workers
The gap
Generic checks miss intent. Senior-style review tests decisions.
Not another static checklist
dryrun is an architecture decision evaluator. Instead of pass/fail lint rules, it pressure-tests proposed designs with cross-verification questions before implementation.
Not a generic scorecard
The evaluation flow includes follow-up interrogation. It asks targeted questions to cross-verify assumptions around reliability, security, cost, and operations.
Not just feedback, but documentation
Finalized decisions are converted into clean ADR drafts your team can review, version, and ship with confidence.
How it works
A senior architecture review workflow, designed for minutes.
Get an initial evaluation
dryrun scores reliability, scalability, security, operability, and cost using explicit evidence and known failure patterns.
Cross-verify decisions
The evaluator asks senior-level follow-ups to test assumptions, reveal weak points, and validate tradeoffs under load and failure.
Generate ADRs
Approved decisions are exported as ADR documents with context, options considered, tradeoffs, and chosen direction.
Evaluation packs
Pressure-test the architecture decisions teams ship.
Multi-tenant SaaS
Tenant isolation, noisy neighbors, schema design, auth boundaries, async billing jobs.
High-traffic CRUD API
Database scaling, cache invalidation, deployment strategy, throttling, and observability.
Event processing pipeline
SQS vs Kinesis, idempotency, poison messages, ordering, replay, and DLQs.
Production incident
Latency spike after deploy. Choose checks, form hypotheses, and avoid tunnel vision.
Threat model review
IAM scope, public exposure, data boundaries, secrets handling, and audit trails.
Terraform review
Explain what your infrastructure implies before cost, reliability, or access mistakes ship.
CLI first
Start in the terminal. Run architecture reviews fast.
Run evaluations directly from your workflow. Generate an evidence-backed report, answer cross-verification prompts, and export ADR drafts without leaving the terminal.
$ npx dryrun
$ dryrun evaluate architecture.md
$ dryrun validate architecture.md
$ dryrun adr generate --from report.json
Early access
Help shape the first architecture evaluator.
Join the early list for the first CLI evaluator with cross-verification and ADR generation.
Waitlist signup is not live yet. We are finishing the backend and opening registrations soon.