RepoDocs AI

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

It helps engineering teams replace scattered docs and fragile template packs with one repository-ready system for structured writing, guided AI usage, validation, examples, publishing, and automation.

Install in 5 Minutes Browse Documentation View System Doc Read Spec Read Summary View Example System See Roadmap
Who It Is For SaaS API Teams

CTOs, engineering leads, DevRel teams, and technical writers working in GitHub-based documentation workflows.

Why It Is Different More Than Templates

Templates, prompts, validation, governance, examples, exports, and automation are packaged together as one documentation system.

Problem

Most teams still document too slowly and too inconsistently.

Documentation drifts from the product, teams use different structures, and AI-generated drafts are hard to trust when there is no shared schema, review process, or validation layer behind them.

Solution

RepoDocs AI gives teams one repo-native documentation workflow.

  • Product and API template packs
  • AI prompts for drafting and review
  • Mermaid starter diagrams
  • Frontmatter, structure, request-body, and deprecation validation
  • Confluence, Google Docs, Notion, and PDF export pipelines
  • Runnable agent, analytics, and knowledge-graph automation
  • Hosted control plane for HTTP-triggered jobs
  • Pull request and GitHub Pages workflows

Example

See a Stripe-style payments docs example, not just templates.

The repository now includes a fuller payments example that covers API overview, endpoints, authentication, structured errors, idempotency, and webhooks so a developer can inspect a believable final output instead of a theoretical starter pack.

How It Works

Problem to published docs in four steps

  • Start with the template library, frontmatter schema, and prompts
  • Generate and refine content with guided AI and review workflows
  • Validate structure, examples, and quality in CI
  • Publish, export, and automate through scripts, workflows, and the control plane

Template Packs

Built around real documentation jobs

Product Documentation

Product overview, feature documentation, user guides, and administrator guides.

API Documentation

API overview and endpoint templates aligned to Markdown-first repo workflows.

Governance

Review checklists and hallucination guardrails for AI-assisted documentation.

Operations

Runbooks and deployment guides for operational readiness and change control.

Architecture

System and integration architecture templates for service design communication.

Roadmap Status

Where the build stands today

Phase 1

Foundation complete

Specification, repository structure, frontmatter schema, and validation guidance are established.

Phase 2

Core packs complete

Product and API templates, prompt packs, and example documents are available in the repository.

Phase 3

Automation baseline shipped

Validation scripts, CI workflows, OpenAPI generation scaffolding, and a publishable site are now implemented.

Quickstart

Create a docs repo, copy the starter system, and start documenting APIs.

A first-time adopter should not have to reverse-engineer how to use RepoDocs AI. The shortest path is to create a docs repository, copy the starter assets, and begin with the API, feature, and governance template packs. 

mkdir company-docs
cd company-docs
cp -R ../repodocs-ai/templates .
cp -R ../repodocs-ai/prompts .
cp -R ../repodocs-ai/diagrams .

Or, from the RepoDocs AI repository root, run one command: 

npm run bootstrap:docs-repo -- ../company-docs
Open install guide View bootstrap source Open docs hub

Feedback

Use RepoDocs AI on a real spec and tell us where it breaks.

The next proof point is external usage. If you try RepoDocs AI on your own OpenAPI file, open the dedicated issue form or join the public GitHub discussion with what worked, what failed, and what still blocks adoption.

Open an issue Open a discussion

Design Partners

Want help applying this to a real API documentation workflow?

Use the design-partner intake in the repository docs to qualify a short pilot conversation around documentation generation, review, and repo setup.

Open the intake questions Review the launch checklist

What Makes It Different

Five subsystems, one documentation system.

  • Template library across product, API, feature, and governance docs
  • Shared frontmatter schema for all Markdown assets
  • Prompt framework for drafting and review
  • Mermaid starters for architecture, sequence, and data flow
  • Validation and publishing automation through scripts and Actions
  • Operations and architecture template packs

Current Status

Foundation, core packs, and automation baseline are already live.

The current repo already ships the product foundation, core document packs, validation workflows, publishing support, documentation sources, and a hosted control-plane baseline. The remaining work is deeper hosting and productization, not first-pass clarity.

Documentation

The live site now includes the source docs.

The published RepoDocs AI site now exposes the core documentation set directly through GitHub Pages instead of limiting visitors to overview pages.

Open docs hub Product spec source System guide source

Published Example

Complete-system sample is now part of the site.

The published site now includes an end-to-end payments example that connects product context, feature docs, API endpoints, governance output, and diagrams in one browsable view.

What to test

Use the live site as a buyer and adopter walkthrough.

Review the roadmap page for spec alignment, the automation page for validation depth, and the example-system page for a full reference implementation.

Next Build

Move from generated documentation to broader publishing and lifecycle depth.

The next milestone is hosted orchestration, broader publishing targets, and deeper lifecycle automation around SDKs and changelogs.