Open-source project
posit-dev/positron avatar
posit-dev/positron

Positron is a VS Code fork with two CI systems and an npm test that refuses to run

Positron, a next-generation data science IDE

4,349 stars206 forksTypeScriptNOASSERTION

At a glance

What is it?
Positron is a data science IDE from Posit PBC built on Code OSS and distributed under the Elastic License 2.0 with date based release numbers. The repository keeps the upstream package identity, adds an ai-lib workspace beside it, and makes the bare npm test fail on purpose.
Who is it for?
Positron suits teams that want a data science editor with a familiar Code OSS base and a documented licence rather than an open source one. It does not suit anyone who needs an MIT repository or a semver train, since the product ships under the Elastic License 2.0 with date numbered builds.
Can I use it commercially?
Check first. The repository uses a licence we do not classify automatically, so read its LICENSE file before any commercial use.
Is it still maintained?
Yes. The repository last received commits 1 day ago.
What is it written in?
Mainly TypeScript, according to GitHub's language statistics.

Answers come from the project's GitHub data, last synced on November 15, 2026, and from our analysis. They are not legal advice.

From daily snapshots

Star history

  • posit-dev/positron4,349 ★ · +74
Star history
Repository2026-09-232026-11-15
posit-dev/positron4.3k4.3k

74 Stars in the last 53 days, measured from the daily GitHub snapshot. Compare with other projects

Editorial analysis

The manifest still names code-oss-dev, with a product label switched on

The package identity in the repository is the upstream one. package.json declares the name code-oss-dev, version 1.138.0, a distro field holding a commit hash, a product-label of positron, an author of Posit Software, PBC, main at ./out/main.js, type set to module, and private set to true. Only the label distinguishes it. That inheritance is the point of the base: Positron is built on Code OSS, so commands, settings and source control behave the way the VS Code documentation describes, and anything documented there applies before any Positron specific behaviour. Two root files record the relationship in practice, a positron.code-workspace for the product and a separate positron-ci-lab.code-workspace for the lab machines. A fork that keeps the upstream manifest is also why a change to versioning or licensing shows up in two different places.

The bare npm test fails on purpose and names the script to run

Running npm test in this repository exits with an error by design. The script prints that a test script from the scripts folder should be run, giving ./scripts/test.sh --run <file> as the example, and then exits 1. The real entry points are all named separately: test:core runs ./scripts/test.sh, test:ext-host runs ./scripts/test-integration.sh, test-browser installs Playwright before running the browser runner, and test:positron runs vitest. Three of them are a bare path with no flags:

code
./scripts/test.sh
./scripts/test-integration.sh
node test/unit/browser/index.js

There is a mocha runner for node units with the tdd interface and a 5000 ms timeout, a vscode-test entry for extensions, a TypeScript check against a vitest tsconfig with skipLibCheck, and a build script test that changes into the build directory first. Nothing about the suite is discoverable from the default command.

MIT in the manifest, Elastic License 2.0 in the documentation

The two licence statements do not match, and the difference is not cosmetic. The manifest declares MIT and private true, which is the ordinary shape for a fork that is never published to a registry. The documentation states that Positron is licensed under the Elastic License 2.0, calls it a source-available license, and points at a licensing page for what that means and why it was chosen. Names are protected separately: Positron and the Positron icon are trademarks of Posit Software, PBC, all rights reserved, so a fork changes neither the licence terms nor the mark.

Two CI systems, plus a self hosted ARM lab

The build runs on more than one system, and the repository keeps both sets of files side by side. A .github directory holds the GitHub side, while a jenkins directory points at the other. Beyond that, the manifest carries two scripts aimed at a lab rather than a laptop: ci-lab-reset and ci-lab-up, both pointing into .devcontainer/ci-arm/, so an ARM machine is created and torn down through scripts checked into the tree. A docker directory and playwright.config.ts cover containerised runs and browser driven tests, and three sample environment files sit at the root, one plain, one for a Jupyter flavour and one for the workbench, each named .env.e2e with a suffix. Release builds for macOS, Windows and Linux desktop are produced from a continuous integration system, which is the part the documentation tells users rather than contributors.

The AI surface is a separate workspace with its own build step

Assistant support is not folded into the main build. There is an ai-lib directory at the root, and the manifest builds it through npm --prefix ai-lib run build with three named workspaces: ai-config, ai-credentials and ai-provider-bridge. The client build itself is split in two as well, a compile-client step that runs gulp compile and a compile-copilot step beside it, joined by a compile script. Testing follows the same separation, with a node script for agent host end to end runs and a second script that produces coverage for the same path. Repository root files point the same way: AGENTS.md and CLAUDE.md, plus .agents, .claude and .zed directories, so instructions are written for several agent tools at once. Version pinning is explicit too, with a .python-version file and an .nvmrc next to them.

Install hooks run TypeScript before npm is finished

npm does not get to run this repository unattended. A preinstall script runs node build/npm/preinstall.ts and a postinstall script runs node build/npm/postinstall.ts, so the install hooks live in TypeScript under the build directory rather than in shell. The same directory holds a cyclic dependency check invoked against the output tree. History is acknowledged rather than hidden: a .git-blame-ignore-revs file lists commits for git blame to skip, a .removed-upstream-dependencies file records what was taken out of the Code OSS base, and .gitmodules shows submodules are in use. Licensing of bundled components is inventoried in cgmanifest.json, cglicenses.json, NOTICE and ThirdPartyNotices.txt.

Release numbers are dates, and the upstream version runs alongside

The published builds are numbered by year and month: 2026.10.0-297 on 2026-10-08, 2026.10.1-2 on 2026-10-19 and 2026.10.2-3 on 2026-10-26. That is a different scheme from the 1.138.0 in package.json, which tracks the Code OSS base instead. Two version lines in one repository is the practical consequence of forking, and it means an upgrade note can refer to either number. The last push to the default branch landed on 2026-11-02, a week after the newest release. Rebuilds are scripted on both platforms through rebuild.sh and rebuild.ps1, with the shared build logic driven from gulpfile.mjs.

Editorial conclusion

Positron suits teams that want a data science editor with a familiar Code OSS base and a documented licence rather than an open source one. It does not suit anyone who needs an MIT repository or a semver train, since the product ships under the Elastic License 2.0 with date numbered builds. Before adopting it, read the licensing page for what source available permits, work out how the upstream Code OSS version and the Positron release number relate to each other, and expect to pick a specific test script rather than run the default one.

Frequently asked questions

how to install positron

Install the latest release from the Positron website download page. The documentation states that release builds are currently produced from a continuous integration system for macOS, Windows and Linux desktop, and the same site covers what to do before installing plus troubleshooting.

how to use positron ide

Positron is built on Code OSS, so basic features such as commands, settings and source control follow the VS Code documentation. The project website carries pre-install information, a troubleshooting section and an FAQ for common questions.

What license is the Positron IDE released under?

The documentation states that Positron is licensed under the Elastic License 2.0, described as a source-available license, with a licensing page explaining the choice. Positron and the Positron icon are trademarks of Posit Software, PBC. The package.json in the repository separately declares MIT and private true.

How do you run the tests in the Positron repository?

The bare npm test deliberately prints an error and exits, telling you to run a script from the scripts folder. Named entries include ./scripts/test.sh for the core suite, ./scripts/test-integration.sh for the extension host, a mocha runner for node units, a Playwright driven browser runner, vscode-test for extensions and vitest for the positron suite.

Official sources

  1. Issues
  2. posit-dev/positron on GitHub
  3. Project website
  4. README
  5. Releases
Add this badge to your README

If you maintain this project, the badge below links readers to this analysis and shows its maintenance status from the daily GitHub snapshot. Paste the markdown into your README; add ?metric=license or ?metric=stars to the image URL for a different field.

Add this badge to your README

markdown
[![Hysen Labs](https://hysenlabs.com/badge/posit-dev-positron.svg)](https://hysenlabs.com/projects/posit-dev-positron)

Star history chart for your README

markdown
[![Star history](https://hysenlabs.com/chart/posit-dev-positron.svg)](https://hysenlabs.com/projects/posit-dev-positron)

Was this useful?