How much time will I actually save with Fazm?

Most users report saving 45-90 minutes per day on repetitive tasks. The biggest time savings come from email management (drafting replies, sorting, archiving), form filling (expense reports, CRM updates, applications), research (comparing products, gathering data from multiple sites), and social media (cross-posting, scheduling, replying). For example, booking a flight that normally takes 10 minutes of clicking through airline sites takes a single voice command with Fazm. Filing an expense report that requires opening a PDF, extracting numbers, and entering them into a spreadsheet becomes one spoken sentence. The more repetitive and multi-step your daily work is, the more time you get back.

What exactly can Fazm control?

Fazm controls everything you can do manually on your Mac. That includes mouse movements and clicks, keyboard typing, browser navigation and form filling, app switching, file management in Finder, and system-level actions. It uses direct browser DOM control for web interactions - meaning it reads and manipulates the actual page structure rather than taking screenshots and guessing where to click. For native macOS apps, Fazm uses accessibility APIs to interact with buttons, menus, text fields, and other UI elements. This works across any application: Chrome, Safari, VS Code, Slack, Figma, Terminal, Mail, Calendar, and more.

How is Fazm different from ChatGPT or Claude?

Chatbots like ChatGPT and Claude give you answers in a conversation window. Fazm takes action directly on your computer. When you say 'reply to Sarah's email,' Fazm opens your email app, finds Sarah's message, writes a contextual reply, and sends it. A chatbot would tell you what to write, but you still have to do all the clicking and typing yourself. Fazm moves your mouse, types on your keyboard, clicks buttons, and navigates between apps. It is a computer agent, not a chat assistant.

How does the memory layer work?

Fazm builds a personal knowledge graph by extracting information from your files, browsing history, conversations, and daily activity. Over time, it learns your contacts (names, emails, relationships), accounts and credentials, preferences (formatting, tone, scheduling habits), and frequently used workflows. In your first week, you might say 'Reply to Sarah's email, she's my cofounder, her email is sarah@acme.com.' By week four, you just say 'Reply to Sarah' because Fazm already knows who she is. By week eight, Fazm can draft the reply before you even ask. All of this data stays locally on your Mac - your knowledge graph never leaves your machine or gets sent to any cloud service.

Is Fazm safe? Can it mess up my computer?

Safety is built into every layer of Fazm. You watch every action in real-time on your screen - Fazm moves your mouse and types visibly, so there is nothing hidden. You can stop any action at any point by pressing a keyboard shortcut. Fazm also shows you what it plans to do before executing destructive actions like deleting files or sending emails, giving you a chance to confirm or cancel. Because Fazm is open source, the code is fully auditable on GitHub. There is no hidden behavior, no background processes making changes you cannot see, and no actions that bypass your screen.

What platforms does Fazm support?

Fazm currently supports macOS, running natively on both Apple Silicon (M1, M2, M3, M4) and Intel Macs. It is built with Swift and uses macOS-specific frameworks like ScreenCaptureKit for screen analysis and Accessibility APIs for app control, which is why macOS is the primary platform. Windows and Linux support are on the roadmap but do not have release dates yet. You can also control your Mac remotely from your phone using Fazm Remote, which gives you voice access to your desktop while you are away from your computer.

How does Fazm voice input work?

One keyboard shortcut activates push-to-talk. Hold the key, speak naturally about what you need done, and release. Fazm transcribes your speech locally using on-device models, so there is no network latency for the voice recognition step. You do not need wake words like 'Hey Siri' or 'OK Google' - just press and talk. You can also type commands into the floating toolbar if you prefer text input. Fazm understands natural language, so you do not need to memorize specific commands. Say 'book me a flight to Tokyo next Thursday' or 'find the cheapest AirPods on Amazon' and Fazm figures out the steps.

Is my data sent to the cloud?

Fazm processes screen analysis and memory entirely on your machine. The personal knowledge graph, file indexing, and screen capture all run locally. The only data that leaves your Mac is the intent - a text description of what you want to accomplish, which is sent to the AI model (Claude or GPT) for action planning. Your actual screen content, file contents, browsing history, and personal knowledge graph never leave your computer. Fazm is also open source, so you can verify exactly what data is transmitted by reading the code on GitHub. For users who need complete air-gapped operation, local model support is planned.

Back to Blog

Accessibility Tree Dumps Overflow LLM Context Windows - How to Fix It

Fazm Team·March 17, 2026·3 min read

accessibility-treecontext-windowllmmacosoptimizationdesktop-agent

Accessibility Tree Dumps Overflow LLM Context Windows

A single accessibility tree dump from a macOS application can easily reach 24KB of raw data. Feed that into an LLM context window a few times during a task, and you have consumed most of your available tokens on structural data that the model barely needs.

This is one of the most common performance killers for desktop AI agents.

The Problem

When an AI agent inspects a macOS application through the Accessibility API, it gets back a hierarchical tree of every UI element - buttons, text fields, labels, scroll views, table cells, and all their attributes. A moderately complex app like Slack or Chrome can produce thousands of nodes.

Most of those nodes are irrelevant to the task at hand. But by default, agents dump the entire tree into their context to "understand" what is on screen. Three or four of these dumps and you have burned through 100KB of context on raw element descriptions.

The Fix

The solution is straightforward: do not put the raw tree in the context window.

Write to /tmp files - dump the full accessibility tree to a temporary file on disk for reference
Return summaries - send back a 5-10 line summary of what is on screen: the key interactive elements, their states, and their positions
Query on demand - when the agent needs details about a specific element, read just that subtree from the file

This approach reduced context consumption from 24KB per inspection to under 500 bytes in practice.

Why This Matters for Agent Performance

Context window space is the scarcest resource for any LLM-powered agent. Every byte of irrelevant data:

Pushes out conversation history and task context
Increases inference cost on token-priced APIs
Reduces the model's ability to reason about the actual task
Slows down response generation

Treating the accessibility tree as a database to query rather than a document to read is the key insight. The agent does not need to see every button on screen - it needs to know which buttons are relevant.

Practical Implementation

For macOS agents using the Accessibility API, write a thin wrapper that captures the full tree, saves it to /tmp, and returns a structured summary. Include element references so the agent can request details about specific nodes when needed.

Fazm is an open source macOS AI agent. Open source on GitHub.

Accessibility Tree Dumps Overflow LLM Context Windows - How to Fix It

Accessibility Tree Dumps Overflow LLM Context Windows

The Problem

The Fix

Why This Matters for Agent Performance

Practical Implementation

More on This Topic

Related Posts

Claude CoWork's Token Limits Hit Different - Why Local Agents Are Better for Big Tasks

The Seven Verbs of Desktop AI - What an Agent Actually Does

What Half a Million Desktop Agent Actions Taught Us About Failure