Promote to production: cache ML models in the service worker - #87
Merged
Conversation
transformers.js fetches Whisper weights from huggingface.co (cross-origin) and its own Cache Storage wasn't reliably persisting for users, so the model re-downloaded on every refresh. Add a service-worker CacheFirst runtimeCaching rule for the HF hosts (incl. cdn-lfs* / *.hf.co redirect CDNs) and cdn.jsdelivr (ORT wasm) — our SW now serves them from cache across refreshes. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
fix(pwa): cache ML model weights so they don't re-download on refresh
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Adds a service-worker CacheFirst rule for HF/jsDelivr so Whisper weights don't re-download on refresh.
🤖 Generated with Claude Code