Swimm, best known for its AI-driven software documentation tools, has announced an extension for GitHub Copilot. The Swimm team says the extension means developers using GitHub Copilot Chat can turn Copilot responses into high-quality, code-coupled documentation, simplifying the process of capturing key insights about users' codebases.

The team adds that the Swimm for GitHub Copilot extension also means that each new document created adds valuable context, making GitHub Copilot “smarter” and improving its ability to generate more accurate, organization-specific responses.

Swimm is a startup that has raised $33.3 million to build AI-driven software documentation tools. The company says the tools are already used by 12,000 teams to auto-document 300 million lines of code.

The new documentation extension auto-generates structured documentation, giving developers a way to automatically take Copilot Chat outputs and transform them into structured, markdown-based documents.

The tool includes continuous documentation sync so that when developers make changes to their code, any related documents are automatically updated to reflect the changes. 

Discussing the launch, Gilad Navot, Co-founder & CPO of Swimm, said

"The Swimm GitHub Copilot extension ensures that the context behind your code—the reasons behind decisions, best practices, and team knowledge—gets documented as you code. This keeps documentation aligned with your code and makes your AI tools more effective."

To use the extension, developers start a conversation with Copilot chat such as a question about their code, such as "What testing frameworks are used in this project?" Copilot then pulls the answers directly from the developer's codebase, providing responses directly within the development workflow.

Once the developer has the information they need, they then use Swimm's command to turn chat outputs into a structured document. Swimm formats everything into markdown, capturing what the code does and why it's set up that way.

Once the document has been created, the developers can enrich it by adding their own insights. Navot says this is where you can include those "unwritten rules" that only the most experienced developers on your team know– like why a certain method was chosen, or how a specific part of the system works.

The automatic updating is a key benefit of the extension, mirroring changes made to the code in the related documentation. Swimm says the tool also helps preserve tribal knowledge, because developers can add business context and team-specific insights directly into documentation, preserving essential knowledge that is often lost.

The Swimm products use deterministic static analysis for code mapping to identify all relevant flows and logical components within the codebase. Relevant context is deterministically retrieved, so that documentation is created grounded in actual code rather than AI-generated assumptions. Only once the data has been retrieved are LLMs introduced to transform the accurately retrieved context into coherent documentation. The Swimm team says this approach means every part of the model's output is anchored to specific parts of your codebase, preventing AI-hallucinations, where the AI-based systems produce highly convincing but incorrect information.

The Swimm Extension for GitHub Copilot is available now. 

More Information

Swimm For GitHub Copilot On GitHub Marketplace

Swimm Website

Related Articles

AI Code Assistants

How To Write Code Documentation

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on Twitter, Facebook or Linkedin.

Comments

Make a Comment or View Existing Comments Using Disqus

or email your comment to: comments@i-programmer.info