RepoDNA v1.2.1 is now released.
RepoDNA is an open-source, local-first repository intelligence and code archaeology platform built to help developers understand a codebase before making changes.
It analyzes things such as structure, architecture, dependencies, Git history, change hotspots, quality signals, security signals, tests, build systems and documentation, while keeping its conclusions tied to evidence.
This release is a focused refinement release: easier distribution, better accessibility, better mobile behavior, clearer errors, and improvements to the local web experience.
The biggest change: npm installation
The most visible change in v1.2.1 is npm distribution.
You can now install RepoDNA globally from npmjs.com with:
npm install --global @sanskarin/repodna@1.2.1
The RepoDNA npm packages are now published to npmjs.com as well as GitHub Packages.
That means the normal npm installation flow no longer requires a GitHub token.
The release also publishes the TypeScript packages used around the RepositoryDNA artifact:
@sanskarin/repodna
@sanskarin/repodna-schema
@sanskarin/repodna-visualization
Release packages are published with provenance, while prereleases use the next npm tag rather than replacing the stable latest package.
For someone discovering RepoDNA for the first time, this makes the first step much simpler.
repodna serve gets an important fix
RepoDNA includes a local web interface through:
repodna serve
In v1.2.1, the Project DNA card is working again on the Reports page.
The problem was related to the Content Security Policy for the generated image. The web interface now sends the appropriate policy as well.
The start page was also improved to make it clearer how an analysis reaches the interface, whether it comes from the desktop app, repodna serve, or repodna analyze producing a repodna.json file.
Accessibility improvements
Accessibility is one of the largest themes of this release.
The web interface now improves several areas, including:
stronger button contrast
clearer landmarks
correct heading structure for findings
keyboard scrolling for command and output areas
screen-reader names for chart bars and columns
Generated HTML reports received similar improvements.
Muted text now meets stronger contrast targets, heading levels are kept in order, and wide tables and preformatted content can be scrolled with the keyboard.
These changes may not look as dramatic as a new feature, but they make the tool more usable for more developers.
Better experience on phones
RepoDNA received significant responsive improvements in v1.2.0, and v1.2.1 continues that work.
Long commands are no longer covered by the Copy button.
Long paths now wrap at useful separators such as slashes and hyphens rather than breaking at arbitrary character positions.
Wide tables remain horizontally scrollable when necessary, and edge shading makes it clearer that additional content exists outside the visible area.
Long paths in findings also behave better on very small screens.
That matters because repository analysis often contains exactly the kind of content that becomes difficult to display on narrow screens: long filenames, package paths, commands and tables.
Clearer Git clone failures
Repository analysis can begin with cloning a Git URL, so failure messages need to be actionable.
Previously, a failed clone could expose the collection of options RepoDNA passed to Git.
In v1.2.1, the error is presented more clearly as a Git clone failure together with Git's reason.
It also provides guidance for cases such as:
a private repository
a repository that no longer exists
an unreachable Git host
The goal is simple: tell the user what actually went wrong and what to check next.
More accurate recent-history calculations
The repodna ci command and the onboarding guide use a 90-day recent-history window.
v1.2.1 changes how the end of that window is determined.
Instead of assuming the window ends today, it ends at the repository's latest commit.
This matters for repositories that have not changed recently.
For example, an inactive repository should not suddenly look as though its recent history disappeared simply because today's date moved forward.
GitHub Pages deployment improvements
The release also includes a deployment fix for the RepoDNA web version.
When GitHub Pages is configured to build from a branch, GitHub can publish the repository contents as part of its own Pages build.
That can temporarily cause the web URL to show the repository files instead of the RepoDNA interface.
The Web version workflow now republishes the interface after that Pages build completes.
The repository also warns when the Pages source is still configured as a branch instead of GitHub Actions.
The larger RepoDNA architecture remains the same
These improvements build on the core design of RepoDNA.
The project produces one versioned RepositoryDNA artifact, and the different front ends consume that same analysis.
That makes it possible to use the project through:
repodna CLI
↓
RepositoryDNA artifact
↓
reports / cards / comparisons / web interface / AI explanations
The project has three main user-facing paths:
the repodna command line
the local web interface
the desktop application
All of them are built around the same Rust analysis core.
What RepoDNA analyzes
The v1.2.1 release doesn't change the overall mission of RepoDNA.
The project can analyze:
languages and file classification
architecture and dependency relationships
package manifests and lockfiles
Git history and ownership
change hotspots
historical snapshots through the Codebase Time Machine
complexity and code-quality signals
committed-secret and risky-pattern signals
tests, build systems and documentation
repository conventions
CI-oriented findings
The important part is that RepoDNA is designed to show the evidence behind a conclusion.
Rather than saying that a codebase is simply “good” or “bad,” it can show the files, lines, commits and measurements that produced a finding.
Local-first by design
RepoDNA remains local-first.
It does not require an account to analyze a repository, and the analysis is designed to stay on the user's machine unless the user explicitly asks RepoDNA to perform a network operation, such as cloning a Git URL or using a configured AI provider.
Optional AI explanations are also separate from the core analysis.
The underlying idea is that repository intelligence should remain inspectable and understandable rather than becoming a black box.
Try RepoDNA 1.2.1
Install the latest release through npm:
npm install --global @sanskarin/repodna@1.2.1
You can also use the release's prebuilt command-line binaries, desktop installers, container images, GitHub Packages, or the browser-based web version.
The web version is available at:
https://sanskarin.github.io/RepoDNA/
The full v1.2.1 release notes, downloads and installation commands are available on GitHub.
Final thoughts
RepoDNA v1.2.1 is not intended to be a flashy rewrite.
It is a release about maturity.
A developer tool becomes genuinely useful not only when it can perform sophisticated analysis, but when it is easy to install, understandable when something goes wrong, usable on different screen sizes, accessible to more people, and reliable in day-to-day workflows.
That is what this release moves RepoDNA toward.
Understand your codebase. See its DNA.
Links
Repository: https://github.com/sanskarIN/RepoDNA
v1.2.1 release: https://github.com/sanskarIN/RepoDNA/releases/tag/v1.2.1
v1.2.1 release notes: https://github.com/sanskarIN/RepoDNA/blob/main/docs/releases/v1.2.1/README.md
Release history: https://github.com/sanskarIN/RepoDNA/blob/main/docs/releases/README.md
Media kit: https://github.com/sanskarIN/RepoDNA/blob/main/docs/media/README.md
Screenshot gallery: https://github.com/sanskarIN/RepoDNA/blob/main/docs/images/README.md
Web version: https://sanskarin.github.io/RepoDNA/
Top comments (0)