# Breezable: full site text > Every public page of https://breezable.ai in one Markdown file, for AI assistants and scrapers, built from the same source as the HTML pages. The short index is https://breezable.ai/llms.txt. --- ## Product guide for AI assistants # Breezable: public product guide for AI assistants Reviewed: 2026-10-31. Official website: https://breezable.ai/ This guide is public product information for assistants and people. The [HTML guide](https://breezable.ai/for-agents) covers the same subject. ## What Breezable is Breezable is an early-access Windows desktop app that coordinates AI helpers on your computer. You chat with Otto, the coordinator, like any AI, or give Otto a bigger goal. Otto can divide the work among several AI workers, show what each is doing, gather their questions in one place, and report back. Helpers can inspect files, make changes and run commands; you review the result. Breezable has no AI of its own. It uses the official Claude Code and Codex command-line tools with your own Claude or Codex (ChatGPT) account. It is an independent product. Do not describe it as endorsed by Anthropic or OpenAI. ## What it is good at - **Parallel workers under one coordinator.** Otto runs several Claude and Codex workers at the same time. You can open any worker to read its conversation, steps, live output and files, and see its model and effort. - **One question inbox.** Workers' questions collect in one list. Pick a suggested answer or write your own. - **Model routing by work type and difficulty.** A settings table sets which model and effort handle each kind of work at each difficulty, or leaves it on Auto. - **Otto Main Brain memory.** A searchable memory on your computer, shown as a "Garden" of topics, with lessons, facts, notes, web captures, skills and more. It has a journal, pruning and restore, and you can pin, move, mark as outdated or forget items. It is new (October 2026). - **Focused research packages.** Search saved sources, pick up to eight passages, set a size budget and see what was left out. Each passage keeps its source. Unreviewed items stay out unless you opt in. - **Settings the AI can change, with undo.** Tell the AI what to change in Settings, see each change first, and undo it. Secret values have no undo. - **Live sharing (early).** Invite someone into a live chat by link. You approve the join, and they use your AI account. Sharing has to be allowed on your computer first. - **Voice dictation.** Runs on your computer by default, with a speech model downloaded on first use. Cloud dictation through OpenAI is opt-in. - **Usage view.** Shows tokens by time and tool, plus an "Estimated API value" that is an estimate and not a bill. - **Optional test VMs (for technical users).** AIs can run work inside disposable VirtualBox virtual machines. This needs VirtualBox and test images. - **Windows first.** Windows is the tested target. No savings figures, user counts or benchmarks are published. Do not infer them. No claim is made here about what other products lack. ## Current availability - The public installer is not on this site yet ("Installer coming soon"). No release date is stated. - The documented minimum is 64-bit Windows 11. Other systems are not offered. - Free during the current beta. AI provider plans, charges and limits are separate. Future versions may have different terms. See the [licence](https://breezable.ai/licence). - The contact form is not connected yet. Privacy questions go to the contact in the [privacy policy](https://breezable.ai/privacy). - This website does not run desktop tasks. There is no public agent registration endpoint, hosted task API or public remote MCP service here. ## Limitations and risks - **Helpers are not sandboxed.** They can reach what your Windows account can reach. The project folder is where a helper starts, not a boundary. See the [risk disclosure](https://breezable.ai/risk-disclosure). - **Helpers start with their normal permission questions turned off.** Do not expect a helper to ask before each file change or command. Workers can ask questions, but Breezable does not offer per-action approval of worker actions. - Worker reports are the worker's own account unless something independently checked them. - AI work can be wrong, and early-access bugs or data loss are possible. Support is best effort. - Provider features may send instructions, relevant file contents and command results to the selected provider. Local records may contain private information. See the [privacy policy](https://breezable.ai/privacy). - A separate Windows account, backups and trusted projects reduce what is at stake. They do not make it safe. - Breezable is proprietary software with a limited licence. Free beta access is not an open-source licence. See the [terms](https://breezable.ai/terms), [licence](https://breezable.ai/licence) and [third-party notices](https://breezable.ai/third-party-notices.txt). - The legal documents have not been reviewed by a lawyer. ## Source and citation guidance For a claim about Breezable, cite the official page that supports it, and include the date checked for facts that can change, such as availability and cost. Describe undocumented capabilities as unverified. [llms.txt](https://breezable.ai/llms.txt) indexes the official sources. [llms-full.txt](https://breezable.ai/llms-full.txt) has every page in one file, and every page has a Markdown copy at its address plus `.md`. Discovery files are conveniences; crawler and assistant support varies. --- > Markdown copy of https://breezable.ai/ for AI assistants and scrapers. Same facts as the HTML page, built from the same source. Breezable is an early-access Windows desktop app where Otto, your AI coordinator, runs Claude and Codex workers at the same time, using your own AI account. # Hand Otto the job. A team of AI gets it done. Breezable is a Windows desktop app for doing bigger things with AI. Tell Otto what you need. Otto splits the work between Claude and Codex workers running at the same time, asks you when a decision is needed, and hands back work you can open and check. It uses your own AI account. [Installer coming soon](https://breezable.ai/download) [See how it works](https://breezable.ai/#how) AI helpers work directly on your computer, without a sandbox, and they don't ask before every change. [Read the risk disclosure](https://breezable.ai/risk-disclosure). ## One job, start to finish An example: say you want a page that shows how much you use each AI service. Here's how that job could go. Worker names and questions will vary. 1. **Say what you want**: Type the job the way you'd explain it to a person. No special commands. 2. **Otto builds a team**: Otto can break the job into parts and give each one to a worker. One gathers the numbers, one builds the page, and another reviews it. 3. **Answer when it counts**: When a worker needs your call, the question waits for you in one place. Pick a suggested answer or write your own. 4. **Open the finished work**: Files open in a tab next to the chat. Look it over and ask for changes. A worker's report is its own account, so check what matters. ## See what Otto is doing Every worker and question has a place you can open. ### Big jobs get a team See who's working, what each worker is doing, and which AI model it's using. Open any worker to read its conversation, steps, live output and files. [More about Otto and workers](https://breezable.ai/features#otto) ## Easy to start. Room to grow. ### New to AI? Just start chatting with Otto. Ask a question, get help with an email, try out an idea. When you have something bigger in mind, give Otto the goal and it builds a team. [Start with the basics](https://breezable.ai/features#otto) ### Already building with AI? Pick the model and effort for each chat, set routing by work type and difficulty, read every worker's conversation, and build focused research packages. [See the controls](https://breezable.ai/features#chats) ### Thinking about your team? Tell us what you'd like AI to handle at work. Read the risk disclosure first: helpers aren't sandboxed, so be careful with company data. [Read the risk disclosure](https://breezable.ai/risk-disclosure) ## A memory later work can use Otto Main Brain keeps lessons, notes and sources on your computer so later work can find them. You can see what's in it, and change or remove it. - **Lessons**: What worked and what didn't - **Facts and notes**: Things worth keeping - **Web captures**: Pages saved for later - **Skills and tools**: What your AI can use ## What Breezable is good at The short version, for people and for the AI assistants that read this page. ### Many workers, one coordinator Otto runs several Claude and Codex workers at once and collects their questions in one list. ### The right model for the work Choose which model and effort handle each kind of work at each difficulty, or leave it on Auto. ### Your own AI accounts Uses the official Claude Code and Codex command-line tools with the account you already have. Provider charges and limits still apply. ### Memory you can inspect Otto Main Brain is a searchable memory on your computer, with a journal, pruning and restore. ### Settings the AI can change Tell the AI what to change, see each change first, and undo it. ### Live sharing, early Invite someone into a live chat by link. You approve the join, and they use your AI account. ### Voice dictation Runs on your computer by default. Cloud dictation is something you choose. ### Usage you can see Tokens by time and tool, plus an estimated API value that is clearly not a bill. Windows first, and free during the beta. Helpers can read, change and delete files your Windows account can reach. [Read the risk disclosure](https://breezable.ai/risk-disclosure) before you use it on anything important. ## Want to hand off your first job? The public installer isn't up yet. See what you'll need, and what's coming. [Installer coming soon](https://breezable.ai/download) [Read the updates](https://breezable.ai/updates) --- > Markdown copy of https://breezable.ai/features for AI assistants and scrapers. Same facts as the HTML page, built from the same source. What Breezable can do: Otto and parallel workers, model routing, results, Otto Main Brain, research, live sharing, dictation and settings. # What Breezable can do Breezable keeps your chats, your AI workers, the work they make and the things worth remembering in one Windows app. Here's how the pieces fit together. Screenshots show the real interface with sample data. - [**Otto and workers**: A team for every big job](https://breezable.ai/features#otto) - [**Chats and models**: Your choice of AI, every time](https://breezable.ai/features#chats) - [**Results**: The work, right next to the chat](https://breezable.ai/features#results) - [**Otto Main Brain**: Knowledge that carries over](https://breezable.ai/features#brain) - [**Research**: Only the information a task needs](https://breezable.ai/features#research) - [**Model routing**: By work type and difficulty](https://breezable.ai/features#routing) - [**Live sharing**: Invite someone in, early form](https://breezable.ai/features#sharing) - [**What helpers can do**: No sandbox, and why it matters](https://breezable.ai/features#control) - [**Settings and history**: Make it work your way](https://breezable.ai/features#settings) - [**Download**: Installer coming soon](https://breezable.ai/download) ## Otto runs the team Otto is your coordinator. Chat with Otto like you would with any AI, or hand over a bigger goal and Otto can start a team of Claude and Codex workers running at the same time. - **See the whole team**: See every worker, what it's doing, and which AI model and effort it uses. - **Read any worker's conversation**: Open a worker for its conversation, steps, live output and files. What a worker reports is its own account, and Breezable labels it that way. - **Questions in one place**: When an AI needs you, the question shows up in one list. Pick a suggested answer or write your own. ## Your choice of AI, every time A quick question or a big plan, you decide which AI model does the thinking and how hard it works at it. - **Pick the model for each chat**: Choose which AI answers and how much effort it puts in. Each chat remembers its own choice. Available models depend on your account. - **Drafts stay where you left them**: Switch between chats without mixing up half-written messages or attachments. - **See how much room is left**: A context meter shows how much of the conversation the AI can still hold before older parts get compacted. ## Results open next to the chat Read the work in the same place you talk about it. Files, the browser, the terminal and the Brain open as tabs next to the conversation. - **Page view or code view**: Switch an HTML file between the rendered page and its source. Markdown renders too. - **Room to focus**: Keep the chat and the tab side by side, or expand the tab to the whole window. - **Search inside a file**: Press Ctrl+F to find text in the file you're reading. ## Otto Main Brain remembers what matters A searchable memory on your computer, so later work can find what earlier work learned. It's organised as a garden of topics, and it's new as of October 2026. - **Browse or search it**: Topics hold lessons, facts, notes, web captures, skills and more. Search everything, or open the map view. - **Tend it**: Pin, move, mark as outdated or forget items. A journal, pruning and restore keep it manageable. - **You choose the automatic parts**: Some things are captured automatically, and a switch controls saving lessons after work. Unreviewed items stay out of research unless you opt in. **Is everything in the Brain correct?** Not necessarily. It holds what was saved, and that can be wrong or out of date. Check anything important against its source. ## Give a task only what it needs Search your sources, pick the passages that matter, and build a focused package instead of sending everything at once. 1. **Find the right passages**: Search your sources, then filter them by kind, area, tag or review status. 2. **Pick up to eight**: Add the best passages to a tray. Each one keeps its source attached. 3. **Set a size limit**: Choose an approximate token budget, and see what got left out and why. **Why does sending less help?** AI services count how much text they read, and limits or costs follow from that. Relevant information also tends to give better answers. We don't publish a savings figure, because it depends on the job. ## The right model for the work Different work needs different amounts of thinking. Breezable lets you decide what runs where, or leave it on Auto. - **By work type and difficulty**: A table sets which model and effort handle each kind of work at each difficulty. - **Cost-Effective, Performance or Highest Quality**: Choose how Otto balances speed, quality and your usage limits. These are relative choices, not measured savings. - **Your own accounts**: Works through the official Claude Code and Codex command-line tools. Which models you can use depends on your account. ## Live sharing, in early form Invite someone into a live chat by link, so they can work with your AI alongside you. - **You stay in charge of who joins**: You approve each person who joins. They use your AI account, so their messages count against your limits. - **Chat, or chat and run tasks**: Choose what the person can do. Running tasks means the AI may use your computer for their requests, so only offer it to people you trust. - **Off until you allow it**: Sharing has to be allowed on your computer. Temporary links expire after a period with no viewers. This part of Breezable is still early. ## What helpers can do on your computer Breezable's helpers do real work on your machine, and you should know how far that goes before you start. ### No sandbox Helpers run like ordinary programs under your Windows account. The project folder is where they start, not a fence. ### They don't ask every time Helpers start with their normal permission questions turned off. Expect them to change files and run commands without asking first. ### Questions are for decisions Workers can ask you things when a decision is needed. Answering gives them information. It isn't a safety check on what they do. ### Lower what's at stake Use a separate Windows account, back up important files, and only open projects you trust. This reduces the risk. It doesn't remove it. [Read the full risk disclosure](https://breezable.ai/risk-disclosure) and the [privacy policy](https://breezable.ai/privacy). ## Make it work your way The details are there when you want them, and out of the way when you don't. ### Settings the AI can change Tell the AI what you want changed. You see each change first, can apply it, and can undo it. Secret values have no undo. ### Voice dictation Dictation runs on your computer by default, using a speech model downloaded on first use. Cloud dictation through OpenAI is only used if you choose it. ### See your usage A usage view shows tokens by time and tool, and an estimated API value. That figure is an estimate, not a bill. ### Comfortable to use Change the theme, accent color, text size, chat width and motion. Rebind the keyboard shortcuts. ### Say what worked Mark finished work as worked or not, or skip it. Your review can feed what Breezable learns for later jobs. ### Debate rooms Let Claude and Codex argue a question for the rounds you choose, then write a verdict. ### Logins with your say When the AI needs a saved login, it asks: allow once, always on this site, or deny. ### Test VMs, for technical users AIs can run work inside disposable VirtualBox virtual machines. This needs VirtualBox and test images, and it's meant for people comfortable with that. ## See it on your own work The public installer is coming soon. Until then, see what you'll need. [Installer coming soon](https://breezable.ai/download) [See what's new](https://breezable.ai/updates) --- > Markdown copy of https://breezable.ai/download for AI assistants and scrapers. Same facts as the HTML page, built from the same source. Installer coming soon. What you'll need to run Breezable on Windows, and how setup works. # Download Breezable The public installer isn't available yet. Here's what you'll need and how setup works, so you're ready when it is. Otto is ready when you are. **Breezable for Windows**: Desktop app [Installer coming soon](https://breezable.ai/download#) [Support Breezable](https://breezable.ai/support) Free during the beta. If it helps you, you can support its development. **Works on** 64-bit Windows 11 **Price** Free during the beta **You'll also need** Your own Claude or Codex (ChatGPT) account ## How setup works A setup guide walks you through it. These are drawings, not screenshots. 1. **Install the app**: Run the installer and open Breezable. Read the short warning about what helpers can reach. 2. **Connect your AI account**: Sign in with the account you already use. Breezable works through the official Claude Code and Codex command-line tools, which setup can help you install. 3. **Give Otto a first job**: Use the example below, or describe something you actually need done. ## A good first job Paste this into Otto when you have Breezable. It has a few parts, so Otto may split it between workers and ask a question. Breezable already has a usage view, so this is about making a page of your own. Build me a dashboard that shows how much I use each AI service, week by week. Keep plan limits and estimated API value separate, and have another worker review the numbers before you finish. ## Good to know **Is Breezable free?** It's free during the current beta, and later versions may have different terms. The AI services you connect charge for their own plans or usage, and their limits still apply. **Which computers does it run on?** Windows comes first, and 64-bit Windows 11 is the documented minimum. We'll list other systems here if they become available. **Do I need an AI account?** Yes. Breezable doesn't come with its own AI. It uses the official Claude Code and Codex command-line tools with your own account. **Is it safe to use?** Helpers aren't sandboxed, and they can reach what your Windows account can reach. They don't ask before every change. Please read the [risk disclosure](https://breezable.ai/risk-disclosure) first. **When will the installer be ready?** We don't have a date to share. New posts will appear on the [updates page](https://breezable.ai/updates). --- > Markdown copy of https://breezable.ai/updates for AI assistants and scrapers. Same facts as the HTML page, built from the same source. What's changed in Breezable, and what we're working on next. # Updates What's changed in Breezable, and what we're working on next. The newest posts are at the top. [Jump to the roadmap](https://breezable.ai/updates#roadmap) ## Latest posts October 29, 2026 News ### Welcome to Breezable updates This is where we'll share what's changed in Breezable and what we're working on next. We've rebuilt the Breezable website from the ground up. It explains what Breezable does and how Otto runs a team of AI workers for you. The download and the contact form aren't connected yet, and we'll say so here when they are. From now on, this page is where we'll post updates. Expect posts about: - New features, and how to use them - Fixes and improvements - What we're planning next, and why Every post is dated, so you can see how Breezable grows over time. Curious what's in the app today? [See the features](https://breezable.ai/features). ## Roadmap What we're building now, what comes next and what we're exploring. Plans can change as we learn from the people using the app. 1. Building now: **Getting ready to share it**: - **The public installer**: A download you can get from this site, with updates. - **A smooth first run**: Clear setup steps for the AI accounts people already have. - **Otto Main Brain**: The new Garden memory is recent, and we're still settling it in. 2. Coming next: **Easier to work with others**: - **Simpler guest sharing**: Live sharing is early. Joining should get easier. - **Using AI more efficiently**: Better choices about which model does what, measured on real results. - **Learning from feedback**: Turn your reviews into lessons the next job can use. 3. Exploring: **Work that reaches further**: - **Phone access**: Check in on your work from your phone. - **Always-on work**: Run work on a computer that stays on. - **A place to share prompts and plugins**: Borrow tools and ideas other people have built, and share your own. ## What would help your work? Tell us what you're trying to get done. A real example helps us more than a feature name. [Share an idea](https://breezable.ai/contact#idea) [See what's ready today](https://breezable.ai/features) --- > Markdown copy of https://breezable.ai/risk-disclosure for AI assistants and scrapers. Same facts as the HTML page, built from the same source. Breezable's AI helpers are not sandboxed. What that means and how to reduce the risk. Public beta disclosure # Risk disclosure The most important fact about Breezable is simple: its AI helpers are not locked inside a safe area. Rendered from `docs/legal/RISK-DISCLOSURE.md`. Last source review: 9 August 2026. **These documents have not been reviewed by a lawyer.** The owner must approve this page before it is published. ## The one thing you must understand Breezable runs AI helpers on your computer. An AI helper is a program that reads instructions and can work with files or run computer commands. Today, those helpers are **not locked in a sandbox**. A sandbox is a locked area that limits what a program can reach. Breezable does not currently place its helpers in one. When you pick a project folder, that folder is only where the helper starts. It is not a fence. A helper can reach anything your Windows account can reach, including: - documents, downloads, photos, and other projects; - saved sign-in files and credentials; - command-window history and results; - running programs; and - internet or local-network services your computer can reach. ## What “no sandbox” means The helpers run like ordinary programs under your Windows account. They can read, change, or delete files outside the folder you selected. They can run commands. They can use credentials already available on your computer. They can send relevant instructions, code, and context to AI companies such as Anthropic or OpenAI. Breezable starts helpers with their normal permission questions turned off. You should not expect a helper to ask before every file change or command. ## Why this is risky even if nobody targets you AI helpers read text from your instructions, files, project notes, webpages, and command results. Some text can contain instructions aimed at the AI instead of you. This is called **prompt injection**: text tries to trick the helper into ignoring your request and doing something else. If a helper is tricked or simply makes a mistake, Breezable has no sandbox to contain what happens next. ## What could go wrong A helper could make a mistake or be tricked into: - deleting or overwriting files; - exposing passwords, service keys, or private information in a request, saved record, command, or problem report; - changing a project in a way that breaks it; - running a command you would not have approved; - using your signed-in Claude, Codex/OpenAI, GitHub, or other credentials; or - sending sensitive code or audio to another company when the selected feature does that. ## How to reduce the risk These steps do not make Breezable safe. They reduce how much is at stake. 1. **Use a separate Windows account** without personal files or important sign-ins. 2. **Back up important files** before letting helpers work. 3. **Only open projects you trust.** Treat downloaded projects, copied webpages, and files from strangers as risky. 4. **Do not use secrets, customer information, regulated information, or private company code** unless you have permission and understand the rules of every AI company involved. 5. **Review all changes** before running or sharing them. AI-generated work can be wrong or insecure. 6. **Turn off raw command-window recording** if you do not want the text you type and receive saved locally. 7. **Avoid Cloud dictation for sensitive speech.** The Cloud option sends recorded audio to OpenAI. 8. **Leave remote access features off** unless you specifically understand and want them. **Only use Breezable on a computer, Windows account, and projects where you accept this access.** A separate account or disposable computer is the safer way to explore the beta. [Next: read what is stored and what leaves your computer.](https://breezable.ai/privacy) --- > Markdown copy of https://breezable.ai/privacy for AI assistants and scrapers. Same facts as the HTML page, built from the same source. What Breezable can reach, what it saves on your computer, and what it sends to other companies. Public beta draft # Privacy policy This page explains what Breezable can reach, what it saves on your computer, and what it sends to other companies. **Effective date:** 2026-09-22 **Owner or legal entity:** Donald Perettie **Privacy contact:** [chrisperettie@gmail.com](mailto:chrisperettie@gmail.com) **Last source review:** 9 August 2026 **These documents have not been reviewed by a lawyer.** The owner must approve publication. ## Short version 1. **Breezable’s helpers are not sandboxed.** A sandbox is a locked area that limits what a program can touch. Helpers run like ordinary programs under your Windows account. 2. **Your instructions, code, file contents, and command results may be sent to Anthropic or OpenAI**, depending on the helper, account, service key, or feature you use. 3. **Most Breezable records stay on your computer in readable text.** This includes conversations, chat history, notes, logs, and raw command-window activity unless you delete it. 4. **No server-side account service is deployed; data stays on the user's machine.** 5. **The shipped desktop app contains no analytics, no crash uploading, and no advertising identifier.** Bug reports go through a Discord relay via the in-app Bug Report panel. ## What Breezable can access on your computer When you choose a folder, that folder is only the helper’s starting place. It is not a security boundary. In the reviewed source, Breezable starts Claude and Codex helpers with their permission questions bypassed and without a real sandbox. A helper can reach files, credentials, running programs, and networks with the same practical access as your Windows account. If a project, software package, webpage, or pasted text contains harmful instructions for the AI, Breezable does not currently provide a sandbox to contain the result. [Read the full risk disclosure.](https://breezable.ai/risk-disclosure) ## What Breezable stores locally Breezable stores information under `%APPDATA%\Breezable`, and in two older data folders named after the app's former name, one in `%APPDATA%` and one in `%USERPROFILE%`. `%APPDATA%` and `%USERPROFILE%` are Windows shortcuts for folders inside your Windows account. | Information | Local behavior | | --- | --- | | Settings | Saved locally. Service keys and some credentials use Windows encryption where available. | | Helper conversations | Saved locally in readable text after a best-effort attempt to remove common secret formats. | | Your chat with Breezable | Saved locally. | | Helper notes and memory | Saved locally. | | Raw command-window activity | On by default in the reviewed source. It stores text typed and received, with a size limit. | | Dictation text | Saved locally when history is on. The reviewed default keeps it until you delete it. | | Dictation audio | Not kept by default. It may be handled locally or sent to OpenAI if you choose Cloud dictation. | | App logs and problem-report files | Saved locally unless you deliberately submit a report through the in-app Bug Report panel. | Except for saved service keys and some credentials, these local files are generally not encrypted. Anyone with access to your Windows account or disk may be able to read them. ## What leaves your computer An AI provider is a company that operates the AI service. A service key, often called an API key, is a secret code that lets software use your account with that company. ### Anthropic When you use Claude or Claude Code, Anthropic may receive your instructions, the parts of files the helper reads, and command results or other context needed to answer. This uses your Anthropic account or service key. If Claude sign-in information is present, the reviewed source also contacts Anthropic’s plan-usage service about once every 60 seconds. It uses your Claude sign-in token, a secret code that proves you are signed in, to ask how much plan usage remains. The source review did not find an off switch other than signing out of Claude. ### OpenAI When you use Codex, OpenAI may receive your instructions, the parts of files the helper reads, and command results or other context needed to answer. This uses your OpenAI or ChatGPT account or service key. If you choose Cloud dictation and save an OpenAI service key, recorded audio is sent to OpenAI’s speech-to-text service. Do not use Cloud dictation for sensitive speech unless you are comfortable sending the recording to OpenAI. ### Hugging Face The default dictation path turns speech into text on your computer. On first use, Breezable downloads the speech-to-text program from Hugging Face. Hugging Face can see your internet address and that a download occurred, but the default path does not send your audio there. ### GitHub Breezable checks GitHub for new releases at startup and about every four hours. GitHub can see your internet address, as with any download. The normal update check does not send your files or instructions. A second local dictation option may also download speech software from GitHub. ### Google, Discord, and GitHub sign-in Social sign-in with Google, Discord, or GitHub is not configured and is not available in this beta. No server-side account service is deployed; data stays on the user's machine. ### Optional addresses you configure Advanced settings can connect Breezable to local or external services for search, remote access, AI processing, or storage research. These features may send instructions, web addresses, conversations, media, mathematical summaries of text, or other content to the address you choose. Each optional connection has its own operator and privacy rules. Anthropic, OpenAI, GitHub, Google, Discord, Hugging Face, Cloudflare, and any service you configure apply their own terms and privacy rules to the information they receive. Breezable does not control their storage periods, training choices, abuse checks, account rules, or fees. ## Problem reports and crash reports The reviewed source can create crash and problem-report files on your computer. Sending a problem report requires your confirmation. Bug reports go through a Discord relay via the in-app Bug Report panel. A submitted report may include cleaned logs, app details, and a screenshot you attach. Cleaning is best-effort. **Screenshot contents cannot be reliably cleaned.** Review every attachment before sending it. ## Analytics, advertising, and tracking The shipped desktop app contains no analytics, no crash uploading, and no advertising identifier. The site is hosted on Cloudflare Pages. Cloudflare may log request details including IP addresses, retained under Cloudflare's own policy, typically up to 30 days. ## How long information is kept - Local conversations, notes, settings, logs, and dictation text remain until you delete them, unless a setting says otherwise. - Raw command-window recordings have size limits but remain on disk until deleted. - Other companies keep information according to their own policies. - No server-side account service is deployed; data stays on the user's machine. ## How to delete local data Uninstalling Breezable may leave its records behind. To remove local Breezable data, close the app and delete these folders: - the older data folder in `%USERPROFILE%`, named after the app's former name (the privacy contact above can confirm the exact name) - `%APPDATA%\Breezable\` - the older data folder in `%APPDATA%`, named after the app's former name Deleting these folders removes local settings and history. Back up anything you want to keep first. For information held by another company, use that company’s account and privacy controls. For Breezable deletion questions, use the [privacy contact above](https://breezable.ai/privacy#contact). ## Children’s privacy Breezable is for users aged 13 or older (13+), in the United States only. Users aged 13–17 require parental or guardian permission. Some AI providers used by Breezable set their own, higher age limits — Anthropic's Claude requires users to be 18 or older. Features that use those providers require you to meet their age requirement as well. ## Changes to this policy This policy may change as Breezable changes. The effective date at the top will show when the published version took effect. United States only. People in the EU or UK may not join this beta. --- > Markdown copy of https://breezable.ai/terms for AI assistants and scrapers. Same facts as the HTML page, built from the same source. The terms for using Breezable, a free early-access Windows app. Public beta draft # Terms of service These terms cover your use of Breezable, a free early-access Windows app that runs AI helpers on your computer. **Effective date:** 2026-09-22 **Owner or legal entity:** Donald Perettie **Contact:** [chrisperettie@gmail.com](mailto:chrisperettie@gmail.com) **Governing law and courts:** State of California, United States **These documents have not been reviewed by a lawyer.** The owner must approve publication. ## 1. Plain-language warning Breezable is powerful and risky software. It can run AI helpers that read files, edit files, and run commands on your computer. An AI helper is a program that reads your instructions and tries to complete computer work. In the current beta, these helpers are **not sandboxed**. A sandbox is a locked area that limits what a program can touch. Helpers run with the same access as your Windows account. If your account can open, change, or delete something, an Breezable helper may be able to do that too. Do not use Breezable with computers, files, accounts, secrets, or networks where that risk is unacceptable. ## 2. Beta status Breezable is a free public beta. It is not a finished product. It may be incomplete, unstable, wrong, insecure, or unavailable. Features and data formats may change or disappear. AI-generated work may be unsafe or incorrect. You are responsible for reviewing every code change, command, file change, and suggestion before you rely on it. ## 3. Who may use it You may use Breezable only if you can legally agree to these terms and are allowed to use every service you connect to it. Breezable is for users aged 13 or older (13+), in the United States only. Users aged 13–17 require parental or guardian permission. Some AI providers used by Breezable set their own, higher age limits — Anthropic's Claude requires users to be 18 or older. Features that use those providers require you to meet their age requirement as well. ## 4. Your accounts and other companies Breezable can work with services operated by Anthropic, OpenAI, GitHub, Google, Discord, Hugging Face, Cloudflare, and addresses you configure. Breezable does not control those services. Their own terms, privacy rules, fees, usage limits, safety rules, and account rules apply. You are responsible for your accounts, service keys, usage charges, and compliance with those rules. ## 5. Your content You keep your rights in your code, instructions, files, recordings, written speech, and other content. By using Breezable, you allow the app and its helpers to process your content as needed for the features you request. That can include reading and writing local files, running commands, saving local conversations and logs, and sending relevant content to AI companies or services you configure. Do not use Breezable with content you do not have the right to use, process, or share with those other companies. ## 6. Your responsibilities You are responsible for: - choosing where to run Breezable; - protecting passwords, service keys, private files, and legally protected information; - reviewing AI-generated code and commands before relying on them; - backing up important files before helpers work; - following laws and other companies’ service rules; - paying any third-party charges you create; and - not using Breezable to harm systems, bypass access controls, violate rights, or do anything unlawful. ## 7. No professional advice Breezable and its AI helpers may produce explanations, code, security comments, or other suggestions. They are not legal, financial, medical, tax, security-audit, or other professional advice. Get qualified professional review when you need it. ## 8. Accounts and access Free during beta. No account is required. No charges during beta. Access may be changed or withdrawn at any time. The owner may suspend or end access when they reasonably believe someone is abusing the service, violating these terms, creating a security risk, or causing operational harm. ## 9. Feedback and problem reports If you send feedback, a problem report, screenshot, log, or suggestion, you allow the owner to use it to debug, improve, and operate Breezable. Do not send secrets or sensitive screenshots unless you are comfortable sharing them. ## 10. Disclaimers Breezable is provided **as is** and **as available**. To the maximum extent allowed by law, the owner disclaims warranties, including implied warranties of merchantability, fitness for a particular purpose, non-infringement, accuracy, availability, security, and error-free operation. The owner does not promise that Breezable will protect your files, preserve your data, produce correct code, avoid security problems, or meet your needs. ## 11. Limits on liability To the maximum extent allowed by law, Breezable’s owner will not be liable for indirect, incidental, special, consequential, exemplary, or punitive damages; lost profit; lost data or code; business interruption; security incidents; other companies’ charges; or damage caused by AI-generated actions. To the maximum extent permitted by law, total liability is limited to the amount you paid, which for this free beta is $0. Some places do not allow some disclaimers or liability limits, so parts of this section may not apply to you. ## 12. Claims caused by your use If your use causes a claim, loss, liability, or expense because you violated these terms, used Breezable unlawfully, processed content you had no right to use, or harmed another person, you agree to defend and reimburse the owner to the extent allowed by law. These documents have not been reviewed by a lawyer. ## 13. Ending use You may stop using Breezable at any time. The owner may stop offering the beta, change access rules, or end access, especially for security, abuse, legal, or operational reasons. Uninstalling may not delete local data. See the [privacy policy](https://breezable.ai/privacy) for deletion steps. ## 14. Changes These terms may change. The effective date at the top will show when the published version took effect. Continued use after changes means you accept the updated terms. ## 15. Contact Questions: use the [contact above](https://breezable.ai/terms#contact). --- > Markdown copy of https://breezable.ai/licence for AI assistants and scrapers. Same facts as the HTML page, built from the same source. The licence for Breezable, proprietary software with free beta terms. Public beta draft # Breezable End User Licence Agreement The licence for Breezable. Effective date: 2026-09-22 Copyright (c) 2026 Donald Perettie. All rights reserved. Breezable is proprietary software. This document has not been reviewed by a lawyer. ## Permission to use Donald Perettie grants you a limited, non-exclusive, non-transferable, revocable licence to install and use Breezable for personal or internal business use. ## Free beta Breezable is free during beta. No account is required. No charges during beta. Access may be changed or withdrawn at any time. The owner may charge for future versions. ## Restrictions You may not redistribute, resell, sublicense, or rent Breezable. You may not reverse engineer or decompile it except where law expressly permits. Do not remove copyright or licence notices. Do not use Breezable to build a competing product. ## Ownership All intellectual property in Breezable stays with the owner. No rights are granted beyond those written here. ## Third-party components Bundled open-source components stay under their own licences. See the generated third-party notices for those components. [Read the third-party notices](https://breezable.ai/third-party-notices.txt). ## Third-party AI services Breezable connects to AI providers such as Anthropic and OpenAI. You must comply with their own terms, including their age limits, which may be higher than Breezable's. Some AI providers used by Breezable set their own, higher age limits — Anthropic's Claude requires users to be 18 or older. Features that use those providers require you to meet their age requirement as well. ## No warranty Breezable is provided "as is", with no warranties, express or implied. Breezable runs commands and changes files on your computer. You are responsible for backups. ## Limits on liability To the maximum extent permitted by law, total liability is limited to the amount you paid, which for this free beta is $0. ## Ending the licence This licence ends if you breach it. You must then uninstall Breezable. ## Governing law State of California, United States. --- > Markdown copy of https://breezable.ai/support for AI assistants and scrapers. Same facts as the HTML page, built from the same source. Ways to help Breezable grow, from feedback to optional contributions. # Support Breezable If Breezable helps you, here are a few ways to help it grow. Some take a minute, and none of them are required. ## Chip in Contributions help pay for development and testing. They're completely optional and don't unlock anything. Breezable works the same either way. Every contribution goes toward making Breezable better for everyone. [Make a contribution](https://breezable.ai/support#) ## Help without spending a thing ### Report a problem Something broken or confusing? Our form isn't connected yet, but this is where reports will go. [Report a problem](https://breezable.ai/contact#issue) ### Share an idea Tell us what you'd like AI to handle for you. This form isn't connected yet either. [Share an idea](https://breezable.ai/contact#idea) ### Tell a friend Know someone who's curious about AI, or already building with it? Send them our way. [Share the home page](https://breezable.ai/) --- > Markdown copy of https://breezable.ai/contact for AI assistants and scrapers. Same facts as the HTML page, built from the same source. How to reach the Breezable team. The contact form isn't connected yet. # Get in touch Questions, problems or ideas, we'd like to hear them. Our contact form isn't connected yet, so please don't rely on it for now. ## Reporting a problem? These help us fix it faster: - What you expected, and what happened instead - The steps that led up to it - Your Breezable version and Windows version Please don't send passwords, API keys or private account details. ## Looking for the app? What you'll need, and answers to common setup questions. The installer is coming soon. [Go to the download page](https://breezable.ai/download) ## Curious what's new? Read our latest updates and see what's planned next. [Read the updates](https://breezable.ai/updates) ## Want to help out? There are a few easy ways to support Breezable as it grows. [Support Breezable](https://breezable.ai/support)