From idea to product
Five steps. Idea to production. No gaps in between.
The full sequence
Most frameworks hand you code and say "good luck." You spend the first week figuring out what order to do things in. What depends on what. Whether you should set up Stripe before or after auth. Whether you need a design system first or can wing it.
We solved that. The build pipeline is a fixed sequence that works for any SaaS product. You open Claude Code in your terminal and run these one at a time:
/discover → /setup → /mvp-spec → /mvp-build (repeat)
Each command feeds into the next. Discovery produces the docs that setup reads. Setup configures the environment that spec generation needs. Specs feed the builder. There are no gaps between steps where you're left wondering what to do.
What each step does
Research your market. Open Claude Code and type /discover. Describe your idea in plain English. Agents go out and study competitors, find the right tools, and validate demand. Then a team of product-thinking agents runs an informed discovery session with you. You end up with seven documents that describe your product completely: what it is, who uses it, what to build, how to price it, what tools it needs.
Set up your environment. Type /setup. It reads those docs and configures everything for you. Creates your database, collects API keys, writes your environment file, and verifies every connection works. Most of it happens automatically.
Generate your specs. Type /mvp-spec. It reads your feature map and writes a specification for every MVP feature. It figures out the build order based on dependencies, groups independent features for parallel work, and produces a roadmap the builder uses to track progress.
Build your features. Type /mvp-build. It picks up the next feature from the roadmap and builds it. Agents read the spec, create a plan, and start working. When it's done, you start a new Claude Code session and type /mvp-build again. One feature per session, until your MVP is complete.
After the MVP
The pipeline doesn't stop at your initial feature set. Once your product is live, you have additional commands for individual features, improvements to existing features, security audits, and performance checks. The same agent system that built your MVP handles everything that comes after.