Moxie Docs: Knowledgebases

Product Introduction

1. Definition: Moxie Docs is a SaaS-based automated codebase documentation and knowledge management platform. It functions as a GitHub-integrated service that continuously analyzes source code repositories to generate, maintain, and serve structured documentation and architectural context.
2. Core Value Proposition: It exists to eliminate documentation drift and knowledge silos in software development by automatically generating living documentation from source code and serving that verified context directly to AI coding agents via the Model Context Protocol (MCP). Its primary value is in reducing onboarding time, improving AI agent accuracy, and ensuring documentation remains synchronized with the codebase without manual maintenance.

Main Features

1. Automated Codebase Documentation Generation: The platform performs a deep structural analysis of connected GitHub repositories, parsing code, imports, and symbols to generate comprehensive documentation. This includes architecture overviews, API references, and coding conventions. It uses static analysis techniques to build a semantic graph of the codebase, which forms the basis for all generated content and drift detection.
2. Continuous Documentation Drift Detection & Friday Cleanup PRs: On every merge to the default branch, Moxie Docs re-indexes affected files and compares the generated documentation state against the new code. It surfaces discrepancies as "drift." For Pro and Team plans, it batches these findings into a single, reviewable "Friday Cleanup" pull request each week, ensuring documentation updates are consolidated and manageable, not disruptive.
3. Model Context Protocol (MCP) Server Integration: Every Moxie Docs workspace includes a dedicated MCP server. This server exposes tools like moxie.get_documentation, moxie.get_conventions, and moxie.get_doc_gaps to compatible AI agents (e.g., Cursor, Claude Code). This allows agents to query live, source-cited context about the codebase instead of performing costly, often inaccurate, full-repository reads, significantly reducing token usage and hallucination.
4. Public-Facing Knowledgebase Hosting (Beta): The service can auto-generate and host public help centers (e.g., your-team.moxiedocs.app) from the same codebase analysis. It separates internal agent notes from public-facing content, requiring manual review and publication for each page, enabling teams to maintain user documentation alongside internal developer documentation.
5. Pull Request Intelligence & Governance: The GitHub App installs checks (Moxie Docs / Description, Documentation, Conventions) that run on pull requests. These checks analyze PR descriptions for completeness, flag potential documentation gaps introduced by the changes, and verify code against project conventions. They are advisory by default but can be configured as required status checks for strict governance.

Problems Solved

1. Pain Point: Documentation Rot and Drift. Traditional wikis and manually maintained docs quickly become outdated as code evolves, leading to misinformation, wasted developer time, and onboarding friction. Moxie Docs directly addresses this by treating documentation as a derived asset from the code, re-validating it on every merge.
2. Target Audience: Engineering Leaders & CTOs seeking to standardize knowledge and reduce tribal knowledge; Staff/Platform Engineers responsible for architectural clarity and onboarding; Developers using AI Coding Agents (Cursor, Claude Code) who need accurate, token-efficient context; Startup and Scale-up Teams experiencing rapid growth where manual documentation cannot keep pace.
3. Use Cases: Accelerating New Hire Onboarding with always-accurate architectural docs; Providing High-Fidelity Context to AI Agents to improve code generation accuracy and reduce costs; Maintaining a Public API or SDK Documentation site that stays in sync with releases; Enforcing Coding Conventions across a team or organization via PR checks; Conducting Codebase Audits through automatically generated architecture maps.

Unique Advantages

1. Differentiation: Unlike static site generators (Docusaurus) or manual wiki platforms (Confluence, GitBook), Moxie Docs generates documentation from code. Unlike other AI doc tools, it focuses on continuous synchronization and MCP-based context delivery rather than one-off generation. Compared to tools like Swimm, it offers a fully hosted, zero-maintenance SaaS model with direct AI agent integration.
2. Key Innovation: The integration of continuous drift detection with AI agent context provisioning via MCP. This creates a closed-loop system: code changes trigger doc updates, and those updates are immediately available to AI agents working on the next change. The "Friday Cleanup" PR is a pragmatic workflow innovation that makes doc maintenance a regular, batch-processed habit instead of a constant interruption.

Frequently Asked Questions (FAQ)

1. How does Moxie Docs handle documentation for forward-looking design docs or RFCs that don't yet have implemented code? Moxie Docs intelligently distinguishes between forward-looking documentation and outdated docs. References in hand-authored documents (like RFCs) to paths or APIs that have never existed in a prior code index are tracked as "never-built" rather than flagged as drift. When code is later implemented that matches those references, the system raises an informational signal for human review, preventing false positives and preserving intentional design documentation.

2. Is my source code secure with Moxie Docs, and how does the GitHub integration handle permissions? Yes, security is a primary design consideration. Moxie Docs installs as a scoped GitHub App, requesting access only to the repositories you explicitly select. It uses encrypted tokens, and its systems are designed to access code solely for the purpose of static analysis to generate documentation and MCP context. No source code is retained for training models or shared outside your workspace.

3. Can Moxie Docs' PR checks block my team from merging code if it finds documentation issues? By default, the PR checks (for description quality, documentation gaps, conventions) are advisory and will not fail the overall status check. However, you can enforce them as a hard gate by adding the specific check names ("Moxie Docs / Documentation", etc.) to your repository's branch protection rules in GitHub, requiring them to pass before a merge can proceed.

4. What programming languages and repository structures are best supported for deep analysis? While Moxie Docs can generate searchable documentation for any GitHub repository, it provides deepest semantic analysis for TypeScript/JavaScript and Python codebases, building detailed import and symbol graphs. It also recognizes and provides structured support for Go, Rust, Ruby, Java, PHP, SQL, and several frontend frameworks. It works with both monorepos and multi-repo setups.

5. How does the MCP server integration actually reduce AI agent costs and improve accuracy? Instead of an AI agent using a high-context window to ingest dozens of files on every prompt—a costly and noisy process—the agent queries the MCP server for specific, pre-digested context. For example, it can ask "What are this project's conventions for API error handling?" and receive a concise, verified answer. This reduces token consumption per task and grounds the agent in actual project patterns, minimizing hallucinations and incorrect API usage.