Repo of the Day
enola-labs/enola
Published: Nov 21, 2026
Open repository ↗Understand, analyze, and govern software systems across repositories, languages, and technologies
Summary
enola turns source code into a local graph of architectural facts: modules, functions, API routes, database tables, message topics, and the relationships between them. It runs as a single binary without requiring an AI model, language server, account, or code upload, and supports repository analysis, change gating, dashboards, coding-agent integration, and cross-repository linking.
What it is useful for
Enola is useful when engineers need more than text search to understand a codebase. Before changing a function, route, or module, its graph can identify the code that depends on it, helping estimate the change’s blast radius.
It can also enforce architecture rules described by a repository. For example, the README shows a new helper in storage importing the delivery layer: the code compiles and tests pass, but an enola check reports the new layer violation. Teams can report such regressions in pull requests with the documented enola-labs/enola-action@v2 GitHub Action.
For systems spread across repositories, enola can connect a frontend request such as /api/v2/orders/{id} to the Go handler serving it, or link a Kafka producer to the service consuming its topic. Coding agents can query the same graph through MCP, allowing them to inspect dependencies before editing and run an architecture check afterward.
The saved graph is also a usable integration format: snapshots consist of plain files containing facts, relationships, findings, and a build receipt, so another tool can load them without rebuilding the analysis.
How engineers can use it
Install the documented binary:
curl -fsSL https://raw.githubusercontent.com/enola-labs/enola/main/install.sh | sh
Start with a read-only report against a checkout:
enola --explain /path/to/your/repo
enola --explain --summary /path/to/your/repo
To generate and inspect a reusable snapshot:
enola --generate .
enola dashboard --open
For change-aware checks, pin the current state, make or pull in changes, and compare:
enola baseline pin
enola check
enola check --fail-on=layers
Agent integration begins with enola install --hooks; the README includes MCP configuration commands for clients such as Claude Code, Codex, Copilot, Cursor, and opencode.
Enola has important boundaries. It models structure rather than runtime behavior, so it does not determine timeouts, retries, or whether messages can be lost. Calls assembled dynamically may remain unresolved, most findings are advisory, and a clean enola check means the change introduced no new findings—not that the repository is already clean.