Auri
Main MascotAuri is our resident alchemist: equal parts wise guide and chaotic builder energy. He represents what StackAlchemist is built for — turning rough ideas into shippable systems, without losing the magic of building.
What is StackAlchemist
Every SaaS starts the same way: an idea, a napkin sketch, and then weeks of grinding through the scaffolding before you write a single line of real business logic. StackAlchemist eliminates that grind. You describe what you're building. We hand you a compiled, production-ready codebase.
The Problem
A typical SaaS scaffolding sprint — entity modeling, API layer, database migrations, auth, Docker setup, folder structure — eats two to four weeks before you ship anything. That's two to four weeks of work that looks the same on every project.
Every developer who has built more than one thing knows this feeling: you've done this before. The pattern is familiar. The implementation is just... tedious. Necessary, but not interesting.
The interesting part is the idea itself. The domain logic. The thing only your product does. That's where you should be spending your time.
The Solution
StackAlchemist takes a natural language description of your SaaS — your entities, relationships, and API endpoints — and generates a complete, compilable source repository in minutes.
Not a template. Not a starter kit. A real codebase, shaped precisely around your schema, with your naming conventions, your business entities, your API surface.
The repository compiles on the first try — or we auto-correct it and try again. That's not a marketing promise. It's a technical guarantee baked into the delivery pipeline.
How It Works
Use Simple Mode to describe your SaaS in plain language, or Advanced Mode to model entities, relationships, and API endpoints with precision. Either way, you're in control of the schema.
Simple Mode — natural language prompt
Advanced Mode — visual entity modeler
Blueprint gives you the architecture document. Boilerplate gives you the full source code. Infrastructure gives you everything, including cloud deployment. Pick the depth you need.
Blueprint · Boilerplate · Infrastructure
Our generation engine applies the Swiss Cheese Method: deterministic Handlebars templates for structure, LLM intelligence injected into the business logic layer. The result is code that looks hand-written, not generated.
Template engine + LLM injection
Before your archive is assembled, the generated code is run through the actual compiler. If it fails, the error output is fed back to the LLM for correction — up to three times. Delivery only happens when the build is green.
dotnet build → auto-correct → retry
You receive a complete ZIP archive: source code, migrations, Docker Compose, environment configuration. Clone it, run docker compose up, and your dev environment is running in under five minutes.
ZIP archive → run locally in minutes
The Method
Most AI code generators produce either too much structure (rigid, one-size-fits-all boilerplate) or too little (hallucinated, inconsistent, non-compilable fragments).
StackAlchemist uses a different approach. The outer structure — file layout, project references, import paths, class signatures — is defined by static Handlebars templates. Deterministic. Reliable. Predictable.
The holes in the cheese — the business logic, domain validation, query implementations — are filled by the LLM, guided by your schema. Intelligence applied precisely where variability is expected. Structure preserved everywhere it matters.
Architecture
Static Handlebars — file structure, class skeletons, import paths
Business logic, domain rules, query implementations
dotnet build validation — structure verified before delivery
3-retry loop with compiler output fed back as context
V1 Stack
Minimal-API endpoint groups with a Dapper repository per entity over Npgsql. One real file each for the model, the repository and the CRUD endpoints.
App Router, TypeScript strict mode, Tailwind CSS. Type-safe API client generated from your schema. Copy .env.example and it runs.
Schema migrations scripted from your entity definitions. UUID primary keys and foreign key constraints modeled from your relationships.
The client library and its environment variables ship wired through Docker Compose. The migration enables row-level security — the policies and the auth flows are yours to write.
One command to spin up the entire stack locally. Database, API, and frontend orchestrated together with proper networking and volume mounts.
A TypeScript stack — VPC, ECS Fargate behind an Application Load Balancer, RDS PostgreSQL. Terraform and Helm equivalents ship beside it.
The architecture is the easy part now.
Mascots
Auri is our resident alchemist: equal parts wise guide and chaotic builder energy. He represents what StackAlchemist is built for — turning rough ideas into shippable systems, without losing the magic of building.
Reto is the specialist behind the Swiss Cheese Method. He keeps structure solid, leaves room for intelligent variation, and reminds us that reliable software is equal parts discipline and craft.