001
MacArena: Benchmarking Computer Use Agents on an Online macOS Environment
In 2nd AIWILD Workshop @ ICML 2026
Advancing Mac-native
intelligence.
MacPaw Research is a research department at MacPaw, focused on LLM efficiency, AI memory, and human–computer interaction — turning findings into the Mac products people use every day. Papers, tech notes, and highlights, published as we go.
§01
Archive
028 / 028002
Help, Don’t Control: Measuring User Emotional Background in AI Assistant Interaction
In Conversational AI: Agency and Identities 2026
003
Composable AI Pipelines: Write Intent, Not Infrastructure
004
GUIrilla: A Scalable Framework for Automated Desktop UI Exploration
In 3rd DATA-FM Workshop @ ICLR 2026
005
Fast-dLLM on MLX: Training-Free Acceleration for Diffusion Language Models on Apple Silicon
006
Do Our Design Principles Ever Really Change, or Do We Just Change the Methods of Achieving Them?
In The International Conference on Computer Graphics, Interaction, Visualization Theory and Applications (GRIVAPP 2026)
Best Poster Award
007
GLiNER2 for Swift: Unified Schema-Based Information Extraction on Apple Silicon
008
Demonstrating Headphone-Sensed, Accessibility-Informed Head Pointing with Snapping for Inclusive Interaction
009
Deep Language Geometry: Constructing a Metric Space from LLM Weights
In RANLP 2025 Recent Advances in Natural Language Processing
010
Screen2AX: Vision-Based Approach for Automatic macOS Accessibility Generation
011
Measure Twice, Cut Once: the Best* Way to Calculate macOS File Hierarchy Size
012
Mousetraps: Reliably distinguishing free apps and trial versions of paid macOS apps on public app marketplaces
013
SwiftEval: Developing a Language-Specific Benchmark for LLM-generated Code Evaluation
In FORGE 2025 The 2nd ACM International Conference on AI Foundation Models and Software Engineering
014
Towards Generating App Feature Descriptions Automatically with LLMs: the Setapp Case Study
In FORGE 2025 The 2nd ACM International Conference on AI Foundation Models and Software Engineering
015
Case Study: Using EEG to Assess Cognitive Load in Infinite Interfaces
In HCI 2025 27TH International Conference on Human-Computer Interaction
016
Datasets: Inside Mac App Store Applications Metadata
017
R+R: State of the Application Sandboxing on macOS: A Differentiated Replication
018
Evaluating a multi-method approach for user emotional dynamics assessment in Software UX
019
Position Paper: Think Globally, React Locally — Bringing Real-time Reference-based Website Phishing Detection on macOS
In STAST'24, 14th International Workshop on Socio-Technical Aspects in Security, Affiliated with the 9th IEEE European Symposium on Security and Privacy
020
Parsing macOS Application UI: Techniques and Tools for Effective Analysis and Automation
021
AI Opportunities in Operating Systems
022
Collecting a Dataset of macOS Apps: Pains, Gains, Lessons Learned. Part 1
023
Application Streaming: Use macOS Apps Remotely
024
Infinite Interfaces
025
Improving macOS Accessibility Features with W1 Chip sensors
026
Passwordless Authentication: Possible Implementation of User Authentication During Registration and Login Using Passkeys
027
Diskoverse. Visualizing macOS Filesystem Using Augmented Reality
028
Static Analysis of macOS Application Bundles for Territorial Affiliation
Partnership
collaborateAcademic & research
University collaborations driving co-authored work, internships, and our long-term talent pipeline.
-
MIT
Internships and joint research with CS students on efficient on-device inference and applied ML.
-
Kyiv-Mohyla Academy
Faculty of Informatics — joint applied-AI research, summer schools, and a steady stream of Kyiv-rooted ML engineers.
-
Ukrainian Catholic University
Faculty of Applied Sciences — joint research, internships, and a steady talent pipeline.
-
KPI Igor Sikorsky Polytechnic
Home of the MacPaw AI Lab "Bilka Space" and a MacPaw-funded scholarship for applied-ML students.
Open to collaboration
Build Mac-native intelligence together.
MacPaw Research partners with universities, research labs, and industry teams on frontier AI — from on-device runtimes and LLM efficiency to AI memory and human–computer interaction. If your work intersects with ours, we'd love to hear from you.
[email protected]
We read every message.