Nicheloom

Market intelligence for builders — see what's gaining traction before it's crowded.

On-device browser agent (Qwen) running locally in Chrome

Details

External ID
46697518
Source
HN
Company
—
Product
On-device browser agent (Qwen) running locally in Chrome
Website domain
github.com
Launched
Jan. 20, 2026
Cohort
—
Upvotes
19
Upvotes percentile
0.647562582345191
Tags
—
Fetched at
Sept. 7, 2026, 9:25 p.m.
Updated at
Sept. 7, 2026, 9:25 p.m.

Description

Demo of LOCAL Browser agent (powered by Web GPU Liquid LFM & Alibaba Qwen models) opening the All in Podcast on Youtube running as a chrome extension.Source: https://github.com/RunanywhereAI/on-device-browser-agentPost: https://www.reddit.com/r/LocalLLaMA/comments/1qh10q9/comment...Geting in the support for web sdk soon, meanwhile have full support for mobile sdks : https://github.com/RunanywhereAI/runanywhere-sdks

Enrichment

Theme
browser automation and scraping for AI
Vertical
Horizontal
Function
Agent / copilot
Audience
Developer
AI stance
AI-native
Project type
Hobby / open-source project
Normalized one-liner
browser agent running locally in chrome
Manually corrected
False

Could you build this?

Partial Building the Chrome extension wrapper is easy, but optimizing and running local quantized LLMs smoothly via WebGPU inside a browser tab requires specialized low-level knowledge of WebGPU runtimes and memory constraints.

What it would actually take: The stack relies on WebGPU, WebAssembly, and libraries like WebLLM / ONNX Runtime Web loading quantized models (e.g., Qwen 0.5B/1.5B) directly into browser VRAM. The difficult parts are managing GPU buffer memory limits inside the Chrome sandbox, preventing main thread freezing during inference, and mapping DOM states into token-efficient visual/textual representations for the on-device agent.

Discussion

3 comments analyzed.

Competitors mentioned: Perplexity, gpt-oss-20b, Ollama

Concerns raised: Potential for malicious use in browser botnets, Unclear complexity limits of tasks

Feature requests: Support for larger local models like 20B+, Distributed computing capabilities for massive models

Competitors

Other products that read as similar to this one — 64 launches clear the similarity bar, closest 8 shown.

Attention rank: #22 of 65 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).

Launched 49 days after the earliest competitor.

Other launches for this product

Same idea, different domain

Nobody's really built a agent / copilot tool for Agriculture yet.