Hugging Face Daily Papers · · 3 min read

Macaron-A2UI: A Model for Generative UI in Personal Agents

Mirrored from Hugging Face Daily Papers for archival readability. Support the source by reading on the original site.

Macaron-A2UI: A Model for Generative UI in Personal Agents</p>\n","updatedAt":"2026-05-26T02:39:36.266Z","author":{"_id":"64303a4484f3ed1ce62a2c5a","avatarUrl":"/avatars/99667daaab49a4e1f2c80bf472cc742b.svg","fullname":"Andrew Chen","name":"anchen1011","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":6,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.698935866355896},"editors":["anchen1011"],"editorAvatarUrls":["/avatars/99667daaab49a4e1f2c80bf472cc742b.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2605.24830","authors":[{"_id":"6a150715b57a1823d5708a3b","name":"Fancy Kong","hidden":false},{"_id":"6a150715b57a1823d5708a3c","name":"Congjie Zheng","hidden":false},{"_id":"6a150715b57a1823d5708a3d","name":"Murphy Zhuang","hidden":false},{"_id":"6a150715b57a1823d5708a3e","name":"Rio Yang","hidden":false},{"_id":"6a150715b57a1823d5708a3f","name":"Sueky Zhang","hidden":false},{"_id":"6a150715b57a1823d5708a40","name":"Hao Fu","hidden":false},{"_id":"6a150715b57a1823d5708a41","name":"Gene Jin","hidden":false},{"_id":"6a150715b57a1823d5708a42","name":"Song Cao","hidden":false},{"_id":"6a150715b57a1823d5708a43","name":"Kaijie Chen","hidden":false},{"_id":"6a150715b57a1823d5708a44","name":"Andrew Chen","hidden":false},{"_id":"6a150715b57a1823d5708a45","user":{"_id":"677aa4d51d4f6bcb2e217c93","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/BPjftZ9rVhaGJfEBRDmLb.png","isPro":false,"fullname":"Xiaoteng Ma","user":"xtma","type":"user","name":"xtma"},"name":"Pony Ma","status":"claimed_verified","statusLastChangedAt":"2026-05-26T07:47:55.349Z","hidden":false}],"publishedAt":"2026-05-24T00:00:00.000Z","submittedOnDailyAt":"2026-05-26T00:00:00.000Z","title":"Macaron-A2UI: A Model for Generative UI in Personal Agents","submittedOnDailyBy":{"_id":"64303a4484f3ed1ce62a2c5a","avatarUrl":"/avatars/99667daaab49a4e1f2c80bf472cc742b.svg","isPro":false,"fullname":"Andrew Chen","user":"anchen1011","type":"user","name":"anchen1011"},"summary":"As personal agents evolve to handle complex, user-centric tasks, static plain-text chat is rapidly becoming a bottleneck. Generative UI emerges as the necessary new interface layer, dynamically synthesizing the right controls, options, and state from the interaction context in real time. We present Macaron-A2UI, a model for Generative UI in personal agents. Our goal is to move beyond text-only interaction by enabling agents to generate natural language together with lightweight, executable UI actions for information collection, preference refinement, confirmation, and multi-goal organization. We build a large-scale Generative UI corpus from heterogeneous dialogue sources, introduce A2UI-Bench for controlled evaluation, and train 30B, 235B and 754B models with parameter-efficient LoRA-based supervised fine-tuning followed by reward-driven reinforcement learning. The best Macaron-A2UI model reaches 75.6 overall on A2UI-Bench without explicit schema hints, surpassing the strongest full-schema frontier baseline. We release the models, benchmark, and evaluation protocol to support future work on Generative UI for personal agents.","upvotes":39,"discussionId":"6a150715b57a1823d5708a46","ai_summary":"Generative UI models enable personal agents to synthesize dynamic interfaces with lightweight executable actions for enhanced interaction beyond text-only formats.","ai_keywords":["Generative UI","personal agents","parameter-efficient LoRA-based supervised fine-tuning","reward-driven reinforcement learning","A2UI-Bench","large-scale corpus"],"organization":{"_id":"69d05e49e99e437a3b18bebc","name":"mindlab-research","fullname":"Mind Lab","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/64303a4484f3ed1ce62a2c5a/0L3sMuvL2JGls3zUKOKYU.jpeg"}},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"64303a4484f3ed1ce62a2c5a","avatarUrl":"/avatars/99667daaab49a4e1f2c80bf472cc742b.svg","isPro":false,"fullname":"Andrew Chen","user":"anchen1011","type":"user"},{"_id":"668775b5c82095d2543b47a3","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/gPgYQAj_WOqTtehiWIa-6.jpeg","isPro":false,"fullname":"NolanHo","user":"NolanHo","type":"user"},{"_id":"67e3e6a8f2cb5de878fea76e","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/uGg83aG0JpiI1cymQlfRq.png","isPro":false,"fullname":"Charles Huang","user":"Charles26","type":"user"},{"_id":"6a03f12eab55afeaedb355c0","avatarUrl":"/avatars/1ea305d92777543b904f1163a7756f6f.svg","isPro":false,"fullname":"Nora Jiang","user":"norrrra","type":"user"},{"_id":"67e926e90afe9169e8dab70f","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/rxew4A5PSaKE07sMeOOit.jpeg","isPro":false,"fullname":"Lam","user":"NoraLam","type":"user"},{"_id":"6621e02cf34ab6caed18e9c6","avatarUrl":"/avatars/15888b2060d1cc56be9fa55fd4b34005.svg","isPro":false,"fullname":"Fanqi Kong","user":"Fancylalala","type":"user"},{"_id":"667bff53f543b957ec305816","avatarUrl":"/avatars/ec60ebe4783d8ebd0dcf2bfc48e2a7ae.svg","isPro":false,"fullname":"yuhuazhou","user":"yuhua-zhou","type":"user"},{"_id":"63f2d83ebe95ed4c9a9fb6f8","avatarUrl":"/avatars/2a2ef80959282c6580b47c32b16748df.svg","isPro":false,"fullname":"LinAnan","user":"Anan168","type":"user"},{"_id":"6a03ef4231a0dd00d0e234a8","avatarUrl":"/avatars/e09298e38cf32c6861a5cfe37db0d825.svg","isPro":false,"fullname":"Cal","user":"VicCale","type":"user"},{"_id":"68a9c660f5f4b7a2a6cb1824","avatarUrl":"/avatars/63be535e1912c463438b7ce0c318db71.svg","isPro":false,"fullname":"caosong","user":"sant1an","type":"user"},{"_id":"64a0154fbfd4d3ffab930448","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/64a0154fbfd4d3ffab930448/lv3tcTlDO-GkjN3kJx5RQ.jpeg","isPro":false,"fullname":"Hongmutian","user":"OriReplication","type":"user"},{"_id":"6969a15c87ebb192324c37e8","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/noauth/qWv0dF3x7ZtKdn0MqrjXW.png","isPro":false,"fullname":"Qihan Liu","user":"liuqh16","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":3,"organization":{"_id":"69d05e49e99e437a3b18bebc","name":"mindlab-research","fullname":"Mind Lab","avatar":"https://cdn-avatars.huggingface.co/v1/production/uploads/64303a4484f3ed1ce62a2c5a/0L3sMuvL2JGls3zUKOKYU.jpeg"},"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2605/2605.24830.md"}">
Papers
arxiv:2605.24830

Macaron-A2UI: A Model for Generative UI in Personal Agents

Published on May 24
· Submitted by
Andrew Chen
on May 26
#3 Paper of the day
Authors:
,
,
,
,
,
,
,
,
,
,

Abstract

Generative UI models enable personal agents to synthesize dynamic interfaces with lightweight executable actions for enhanced interaction beyond text-only formats.

AI-generated summary

As personal agents evolve to handle complex, user-centric tasks, static plain-text chat is rapidly becoming a bottleneck. Generative UI emerges as the necessary new interface layer, dynamically synthesizing the right controls, options, and state from the interaction context in real time. We present Macaron-A2UI, a model for Generative UI in personal agents. Our goal is to move beyond text-only interaction by enabling agents to generate natural language together with lightweight, executable UI actions for information collection, preference refinement, confirmation, and multi-goal organization. We build a large-scale Generative UI corpus from heterogeneous dialogue sources, introduce A2UI-Bench for controlled evaluation, and train 30B, 235B and 754B models with parameter-efficient LoRA-based supervised fine-tuning followed by reward-driven reinforcement learning. The best Macaron-A2UI model reaches 75.6 overall on A2UI-Bench without explicit schema hints, surpassing the strongest full-schema frontier baseline. We release the models, benchmark, and evaluation protocol to support future work on Generative UI for personal agents.

Community

Paper submitter about 5 hours ago

Macaron-A2UI: A Model for Generative UI in Personal Agents

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.
Tap or paste here to upload images

· Sign up or log in to comment

Get this paper in your agent:

hf papers read 2605.24830
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 3

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2605.24830 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2605.24830 in a Space README.md to link it from this page.

Collections including this paper 1

Discussion (0)

Sign in to join the discussion. Free account, 30 seconds — email code or GitHub.

Sign in →

No comments yet. Sign in and be the first to say something.

More from Hugging Face Daily Papers