Control your X/Twitter feed using a small on-device LLM
Details
- External ID
- 47706293
- Source
- HN
- Company
- —
- Product
- Control your X/Twitter feed using a small on-device LLM
- Website domain
- imbue.com
- Launched
- April 9, 2026
- Cohort
- —
- Upvotes
- 15
- Upvotes percentile
- 0.7062982005141388
- Tags
- —
- Fetched at
- Sept. 7, 2026, 9:26 p.m.
- Updated at
- Sept. 7, 2026, 9:26 p.m.
Description
We built a Chrome extension and iOS app that filters Twitter's feed using Qwen3.5-4B for contextual matching. You describe what you don't want in plain language—it removes posts that match semantically, not by keyword.What surprised us was that because Twitter's ranking algorithm adapts based on what you engage with, consistent filtering starts reshaping the recommendations over time. You're implicitly signaling preferences to the algorithm. For some of us it "healed" our feed.Currently running inference from our own servers with an experimental on-device option, and we're working on fully on-device execution to remove that dependency. Latency is acceptable on most hardware but not great on older machines. No data collection; everything except the model call runs locally.It doesn't work perfectly (figurative language trips it up) but it's meaningfully better than muting keywords and we use it ourselves every day.Also promising how local / open models can now start giving us more control over the algorithmic agents in our lives, because capability density is improving.
Enrichment
- Theme
- alternative social platforms and feed readers
- Vertical
- Media & entertainment
- Function
- Agent / copilot
- Audience
- B2C
- AI stance
- AI-native
- Project type
- Hobby / open-source project
- Normalized one-liner
- local llm for twitter feed curation
- Manually corrected
- False
Could you build this?
Partial The browser extension and feed-scraping logic are easily vibe-coded, but running a 4B parameter model efficiently on-device across browsers and mobile requires heavy optimization.
What it would actually take: The stack needs a Chrome extension and iOS app using WebGPU (via WebLLM/Transformers.js) or CoreML/Metal bindings to execute quantized model weights locally. The hard parts include 4-bit model quantization, minimizing RAM usage to avoid browser and mobile OS memory kills, and managing batch latency to prevent DOM stutter. This requires engineers experienced in on-device model optimization and native mobile GPU acceleration.
Discussion
3 comments analyzed.
Competitors mentioned: Twitter, WebLLM
Concerns raised: Battery life on mobile devices, Performance optimization for smooth scrolling, Model optimization not yet production-ready
Feature requests: Better personalization/tailoring of content feed, Fully on-device processing without external servers, Support for more modern multimodal models
Competitors
Other products that read as similar to this one — 56 launches clear the similarity bar, closest 8 shown.
Attention rank: #18 of 57 (itself plus its competitors, highest first — normalized so YC and Product Hunt are compared fairly).
Launched 160 days after the earliest competitor.
- tweet.md · ph · 2026-05-25 · 239 upvotes · similarity 0.44
- twitter-x-delete-suite · github · 2026-09-20 · 26 upvotes · similarity 0.43
- Better Twitter · ph · 2026-09-17 · 8 upvotes · similarity 0.42
- SeenX · github · 2026-09-26 · 8 upvotes · similarity 0.40
- FeedSense · hn · 2026-04-10 · 7 upvotes · similarity 0.40
- Oku · hn · 2026-03-19 · 20 upvotes · similarity 0.39
- Bulkmark · ph · 2026-05-23 · 151 upvotes · similarity 0.38
- xpaiming · ph · 2026-09-21 · 1 upvotes · similarity 0.38
Other launches for this product
- No other launches for this product.
Same idea, different domain
Nobody's really built a agent / copilot tool for Agriculture yet.