Add Firefox AI integration plan: wllama as recommended provider#44
Open
nitzanpap wants to merge 3 commits into
Open
Add Firefox AI integration plan: wllama as recommended provider#44nitzanpap wants to merge 3 commits into
nitzanpap wants to merge 3 commits into
Conversation
Research and plan for bringing AI features to Firefox, which currently has them disabled due to WebLLM's 4.5MB tokenizer exceeding Firefox's 5MB per-file add-on store limit. Evaluated browser.trial.ml (Firefox's built-in trial API) and wllama (WebAssembly llama.cpp bindings). Recommends wllama because it provides chat completions API matching our existing architecture, works on Firefox stable (not just Nightly), and Mozilla has explicitly enabled it for extensions (Bug 1976704). https://claude.ai/code/session_01QwsQd35r81rW7X2PyjBYK6
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Implement on-device AI features for Firefox using wllama (WebAssembly llama.cpp bindings), replacing the previously disabled WebLLM which exceeded Firefox's 5MB per-file add-on store limit. Changes: - Add WllamaProvider implementing AiProviderInterface with chat completions via GGUF models (Qwen2.5-3B, SmolLM2-1.7B, Phi-3.5) - Add @wllama/wllama dependency with bundled WASM files - Update AiService to route to correct provider based on browser (WebLLM/WebGPU for Chrome, wllama/WASM for Firefox) - Add wasm-unsafe-eval CSP directive to Firefox manifest - Add Vite plugins: stub wllama for Chrome, stub WebLLM for Firefox, copy WASM assets for Firefox builds - Remove Firefox AI block from popup and sidebar UI - Make vramRequiredMb optional in AiModelConfig (CPU-only provider) - Auto-detect browser and set correct AI provider on startup - Add comprehensive WllamaProvider tests (27 tests) - Add wllama provider routing tests to AiService (5 tests) All 663 tests pass, both browser builds succeed. https://claude.ai/code/session_01QwsQd35r81rW7X2PyjBYK6
|
any update about this? |
Owner
Author
Hey, Not yet, I've been a bit busy. Hopefully I'll find the time to work on this :) |
Resolve merge conflict in tests/AiService.test.ts by keeping master's vi.spyOn approach and adapting wllama provider routing tests to use spies on the actual wllamaProvider singleton. https://claude.ai/code/session_01QwsQd35r81rW7X2PyjBYK6
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
3 participants
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.
Research and plan for bringing AI features to Firefox, which currently
has them disabled due to WebLLM's 4.5MB tokenizer exceeding Firefox's
5MB per-file add-on store limit.
Evaluated browser.trial.ml (Firefox's built-in trial API) and wllama
(WebAssembly llama.cpp bindings). Recommends wllama because it provides
chat completions API matching our existing architecture, works on
Firefox stable (not just Nightly), and Mozilla has explicitly enabled
it for extensions (Bug 1976704).
https://claude.ai/code/session_01QwsQd35r81rW7X2PyjBYK6