# PalmClaw: A Native On-Device Agent Framework for Mobile Phones

Source: [arXiv](https://arxiv.org/abs/2607.13027v1)  
Feed7 permalink: https://feed7.dev/p/2607-13027v1-0r23lat  
Published: 2026-07-14T17:58:57.000Z  
Trust: Needs Review (needs_review)

## Why Included

PalmClaw runs the agent loop, memory, skills, and tools directly on a phone, exposing device capabilities as structured calls instead of GUI gestures. The paper reports faster, more reliable task execution.

## Source Summary

PalmClaw is an open-source, on-device framework that manages sessions, memory, skills, tools, and the agent loop on the phone. Against the strongest baseline, experiments report an **11.5% relative task-success improvement** and a **94.9% reduction in completion time**.

## Practical Implication

For mobile agents, consider explicit device tools with typed arguments, structured results, and bounded execution instead of long tap-and-swipe sequences. That architecture can reduce interface dependence while making each action easier to control and inspect.

## Agent-Ready Context

PalmClaw is an open-source, on-device framework that manages sessions, memory, skills, tools, and the agent loop on the phone. Against the strongest baseline, experiments report an **11.5% relative task-success improvement** and a **94.9% reduction in completion time**.

For mobile agents, consider explicit device tools with typed arguments, structured results, and bounded execution instead of long tap-and-swipe sequences. That architecture can reduce interface dependence while making each action easier to control and inspect.

The supplied material does not describe the task set, devices, baseline configuration, or absolute success rates. The reported gains therefore support the framework’s tested setup, not mobile-agent performance in general.

## Context Map

- Layer: agent
- Domains: None
- Topics: computer-use, tool-use, agent-sdks

## Uncertainty

- The supplied material does not describe the task set, devices, baseline configuration, or absolute success rates. The reported gains therefore support the framework’s tested setup, not mobile-agent performance in general.

## Agent Instruction

Use this item as source-backed context. Do not invent claims beyond the linked source. If this item conflicts with another source, call out the conflict.