Spec Summary

RepoDocs AI is an AI-prompt-powered docs-as-code documentation system for SaaS teams building APIs.

It gives engineering teams a repository-ready documentation system with templates, prompt packs, validation guardrails, examples, and automation so they can create clearer docs faster and keep them aligned with product and API changes.

Problem

Normal documentation templates leave too much work undone.

  • Teams document features and APIs inconsistently across repositories
  • Manual documentation is slow and often slips behind releases
  • AI-generated docs are unreliable without structure and review guardrails
  • Most template packs stop at layout and do not help with validation or workflow automation

Solution

RepoDocs AI turns documentation into a repeatable repo-native system.

  • Reusable Markdown templates across product, feature, API, governance, operations, and architecture docs
  • Prompt packs for drafting and review instead of ad hoc AI usage
  • Shared frontmatter schema and validation scripts for consistency
  • Examples, CI workflows, publishing support, exports, and a hosted control plane for automation

Example

What the system looks like in practice

  • A product team starts with a structured repo instead of blank Markdown files
  • An engineer feeds OpenAPI or product inputs into guided prompts
  • The team fills consistent templates, reviews with checklists, and validates structure automatically
  • The published site then shows a complete example system instead of isolated doc fragments

How It Works

Problem to published docs in four steps

  • 1. Start with the template library, frontmatter schema, and prompts
  • 2. Generate and refine content with guided AI and review workflows
  • 3. Validate docs quality, OpenAPI examples, versions, and structure in CI
  • 4. Publish, export, and automate updates through scripts, workflows, and the control plane

Why It Is Different

More than normal documentation templates

  • Templates are combined with AI prompt frameworks, not used in isolation
  • Validation and governance are built in, not left to manual review only
  • Examples and site publishing are part of the system, not separate follow-up work
  • Automation, exports, and a hosted runtime are included for operational use

Current Product Status

Foundation, core packs, and automation baseline are already shipped

  • Phase 1: foundation complete
  • Phase 2: core template packs complete
  • Phase 3: automation, export, starter-pack, and hosted-control-plane baseline shipped
  • Remaining work is mostly deeper hosting, tenancy, and productization

Next Click

Use the site in the same order a new buyer would.

Read the system page for the product walkthrough, open the example system to see how the documentation model works in practice, and use the Docs hub for the full source documentation set. The repository copy of this summary still lives at docs/roadmap-spec-summary.md.