npm install babylonpress-ktx2-encoder
Just published to NPM. The old version was used with well-known GLB Optimizer. The new greatly improved encoder will be used in GLB Optimizer V2 (will be published soon) and is already used in the brilliant AO Lightmapper (dont miss free License giveaway)!
What has improved
Faster batch encoding without blocking the UI
Performance is one of the biggest improvements. The browser encoder can now run directly, in a dedicated worker, or through a reusable worker pool. For GLBs containing many textures, the pool processes several encodes concurrently and can be much faster than sequential main-thread batch encoding.
Worker and pool modes move the expensive Basis compression work off the browser’s UI thread. The application remains responsive while textures are encoded instead of freezing during a long conversion. A single worker is primarily a responsiveness improvement; the largest throughput gains come from using a tuned worker pool for batch jobs. Encoding requests also support cancellation and progress lifecycle events.
Basis JavaScript and WASM assets are resolved automatically in the browser build. The Node.js loader reads the packaged WASM binary directly, avoiding fragile relative-URL loading. Input validation reports invalid quality ranges, unsupported option combinations, malformed cubemaps, and missing Node decoders before expensive encoding begins. Output allocation can grow when necessary instead of relying on a single fixed-size buffer.
The glTF-Transform integration has also been hardened. Converted external textures receive .ktx2 URIs, slot and name filters behave consistently, conversion failures reject the transform instead of being silently hidden, and KHR_texture_basisu is marked as required when textures are converted.
Finally, the package includes Node and browser test suites, real Chrome smoke tests, benchmark pages, an interactive image-conversion demo, a runnable GLB example, and a consumer-style package smoke test. These additions make the encoder easier to validate, integrate, and publish with confidence.
The result is the extremely useful Basis Universal encoder, adapted for fast responsive web applications, repeatable Node.js pipelines, and practical GLB/GLTF optimization.