Sage AI

Sage AI is a generative AI tool that automates the creation of documentation, reference guides, and walkthroughs for codebases. This software streamlines the onboarding process for new engineers and helps maintain expertise in complex or aging systems, significantly reducing the time spent on these tasks.

Key features:

- Autogenerated Documentation: Produces detailed documentation at file and directory levels, including overviews of functionalities, dependencies, and project context, linked directly to source code and automatically updated with changes.
- Code Walkthroughs and Reference Guides: Generates guides to help engineers quickly understand codebase structure and functionality, particularly useful for new team members or when working on unfamiliar code sections.
- Documentation Freshness Enforcement: Integrates with knowledge bases and source control systems to monitor and enforce up-to-date documentation, surfacing potential staleness caused by pull requests and incorporating incentives for documentation maintenance into the developer workflow.
- Symbol-Level Contextual Knowledge Base: Constructs a symbolic graph representation of codebases, repeatedly traversed to enrich the knowledge base, clarifying large, complex, or aging systems.
- Contextual Chat: Allows users to chat with the AI for context-specific information from both source code and knowledge base, with customizable starting context and codebase traversal specifications.

How it works:

1. Users interact with Sage AI through a web app or GitHub app.
2. The tool integrates with source control systems to take knowledge base snapshots associated with commits.
3. Sage incrementally adjusts and maintains documentation freshness based on these snapshots.
4. Users can customize query context to receive detailed, context-specific responses.

Integrations:

GitHub

Use of AI:

Sage AI utilizes generative artificial intelligence to create and maintain documentation and its knowledge base. It constructs a symbolic graph representation of codebases and uses this to generate detailed, context-specific documentation.

AI foundation model:

The specific AI foundation model or Large Language Model (LLM) used by Sage AI is not explicitly mentioned. However, the tool goes beyond simply feeding source files into LLMs by constructing and traversing a symbolic graph representation.

Target users:

- Engineering teams working with large, complex, or aging codebases
- Organizations aiming to streamline the onboarding process for new engineers
- Teams seeking to maintain up-to-date documentation with minimal manual effort

How to access:

Sage AI is available as a web app and a GitHub app. It is a proprietary tool developed and maintained by Emblem Software, Inc.

Example Use Cases:

- Onboarding New Engineers: Sage AI can reduce the calibration time for new engineers by nearly half through autogenerated documentation and walkthroughs.
- Maintaining Documentation: By integrating with pull requests, Sage ensures documentation remains current, reducing time spent on manual updates.
- Understanding Complex Codebases: The symbol-level knowledge base aids engineers in quickly comprehending the structure and dependencies within large or aging codebases.