Skip to main content
Waterline indexes your repository at the function and class level. This lets it pinpoint exactly which part of your codebase implements a given acceptance criterion — not just which file, but which function.

​
What gets indexed

For each source file, Waterline extracts individually meaningful units of code: functions, methods, classes, and their documentation. Each one is analyzed in context and stored with a semantic understanding of what it does — not just its name. Binary files, lock files, build output, and dependency directories (node_modules, .git) are excluded automatically.

​
Initial indexing

When you first connect a repository, Waterline crawls every source file in parallel and builds the index. For most repositories, this completes in 5–15 minutes. Progress is shown in the dashboard.

​
Staying in sync

Waterline registers a GitHub webhook on every connected repository. Each push to any branch triggers an automatic incremental update — only changed files are re-processed. For most pushes, the index is updated within a few minutes, without any manual action.

​
Your code is not stored

Waterline reads your source files to build a semantic understanding of what each function and class does. It does not store your raw source code. What’s retained is the semantic index — summaries and embeddings — not the code itself.

​
Supported languages

Waterline supports all major languages used in production codebases, including TypeScript, JavaScript, Python, Go, Ruby, Java, Kotlin, Rust, and more. If you’re unsure whether your stack is covered, contact us.