Why it matters
- Documentation debt is universal — Stenography addresses it systematically rather than relying on developers to manually write comments.
- VS Code integration with keyboard shortcuts makes documentation generation as easy as saving a file — no workflow disruption.
- Multi-language docstring format support (Google, NumPy, reST, JSDoc) means generated docs are immediately compatible with existing codebases.
- Batch API enables retroactive documentation of legacy codebases — a task too tedious to do manually.
Key capabilities
- Docstring generation: Python (Google, NumPy, reStructuredText), JavaScript (JSDoc), TypeScript, and more.
- Inline comments: Explain complex logic blocks, algorithms, and non-obvious code decisions.
- VS Code extension: Keyboard shortcut to document selected function or file without leaving the editor.
- API: REST API for batch documentation generation — pipe entire directories for mass documentation.
- Plain English summaries: Generate README-style explanations of what modules and classes do.
- Multi-language: Python, JavaScript, TypeScript, Rust, Go, C/C++, Java, Ruby, PHP.
- Codebase indexing: Understand function relationships and dependencies for more accurate documentation.
Technical notes
- IDE integration: VS Code extension; API for CI/CD and batch processing
- Languages: Python, JavaScript, TypeScript, Rust, Go, C/C++, Java, Ruby, PHP
- Docstring formats: Google, NumPy, reStructuredText (Python); JSDoc (JavaScript/TypeScript)
- Pricing: Paid subscription (per seat); enterprise API pricing available
- Founded: 2021 by Bram Adams; United States
Ideal for
- Teams with legacy codebases that lack documentation and need to bring them up to standard efficiently.
- Engineering organizations that want documentation generation as part of CI/CD — auto-document new code on commit.
- Individual developers who write code quickly but consistently underdocument due to time pressure.
Not ideal for
- Teams who primarily want overall architectural documentation — Stenography documents code, not system design.
- Projects where AI-generated docstrings could introduce inaccuracies in safety-critical or complex algorithm documentation — human review required.
- Developers who already use Cursor or GitHub Copilot with documentation generation capabilities built in.
See also
- Mintlify — Documentation site builder with AI writer; generates developer docs sites, not code comments.
- Cursor — Full AI IDE; can generate docstrings but requires manual prompting per function.
- GitHub Copilot — General coding AI that can generate docstrings inline as you type.