An 8,435-star repo reduced to a redirect notice
This was the WebGL-accelerated core of TensorFlow.js, bringing linear algebra and automatic differentiation to TypeScript before the code was absorbed into a monorepo.

What it does This was the low-level engine behind TensorFlow.js: tensor operations, linear algebra, and automatic differentiation, all WebGL-accelerated and written in TypeScript. It let browsers run and train neural networks without plugins. Now the code is frozen as a historical archive.
The interesting bit
The README is essentially a tombstone. There are no docs, no build instructions, no benchmarks—just a notice that work continues inside the tfjs-core folder of the tensorflow/tfjs monorepo. It is a remarkably concise end to a once-active standalone project.
Key highlights
- Archived; all future PRs must go to the
tensorflow/tfjsmonorepo. - History and contributions were preserved during the migration.
- Repository description still advertises WebGL-accelerated ML, linear algebra, and automatic differentiation for JavaScript.
- 8,435 stars remain as a measure of its earlier standalone popularity.
Caveats
- Active development has ceased here.
- The README offers no technical detail beyond the redirect.
Verdict Useful only as archaeological evidence for how TensorFlow.js was structured before monorepo consolidation. Active developers and contributors should ignore this and go straight to the main repository.
Frequently asked
- What is tensorflow/tfjs-core?
- This was the WebGL-accelerated core of TensorFlow.js, bringing linear algebra and automatic differentiation to TypeScript before the code was absorbed into a monorepo.
- Is tfjs-core open source?
- Yes — tensorflow/tfjs-core is open source, released under the Apache-2.0 license.
- What language is tfjs-core written in?
- tensorflow/tfjs-core is primarily written in TypeScript.
- How popular is tfjs-core?
- tensorflow/tfjs-core has 8.4k stars on GitHub.
- Where can I find tfjs-core?
- tensorflow/tfjs-core is on GitHub at https://github.com/tensorflow/tfjs-core.