Meet 
Mast
An AI teammate you can give real work to. Mast signs in to your tools, works around the clock, and comes back with finished work.
Acme follow-upSigning in to Attio to pull the deal history.
- Attio deal history pulled · 14 touches
- Gmail draft sent to Rob, cc Priya
- Calendar hold added for Thursday
Message Mast like a teammate
Give Mast tasks the way you would a coworker: on the web, in Slack, or by email on its own address. It keeps context on how you work and checks in when it needs your approval.
A teammate who actually does the work
Mast is not a chatbot you ask things. It picks up the task, does it in your tools, and brings back the finished thing for review.
Mast signs in to your tools
Connect Gmail, Slack, Attio, QuickBooks and more through OAuth scopes you approve. Mast uses them the way you would.
Show Mast how it's done
Handle a task a few times and Mast offers to take it over. You stay in the approval loop until you decide you don't need to be.
Mastnoticed a patternYou've sent this follow-up three times this week. Want me to draft these in your voice from now on?
It works while you sleep
Mast wakes on schedules and triggers: a new email, a calendar event, a CRM update. The work is done before you sit down.
- 6:10 AMInbox triaged
- 6:40 AMAds report drafted
- 7:05 AMBooks reconciled
Nothing ships without you
Mast proposes, you approve. Sends, record updates, and meetings wait for your OK until you loosen the leash.
Draft reply to Acme is ready. Review before I send?
Give Mast a job
Inbox at zero by 9 AM. Drafts replies in your voice, books meetings, reschedules conflicts, and preps you for every call.
Mast lives where your team works
Mast gets its own email address and shows up in every channel your team uses. One memory across all of them. It remembers the thread no matter where it started.
Pick the brain behind Mast
Bring your own model, or connect the ChatGPT subscription you already pay for. Mast can switch models per task: cheap for triage, sharp for memos.
Sign in with ChatGPT. Routes through your Plus, Pro, or Team credits so you don't pay twice.
Bring your Anthropic key. Sonnet, Haiku, Opus.
Bring your OpenAI key. GPT-5 and reasoning models.
Bring your Google AI key. Gemini 2.5 Pro and Flash.
Bring your xAI key. Grok-4 reasoning.
Locked down before it wakes up
Mast runs in its own isolated cloud VM, authenticates through OAuth scopes you grant, and only touches the tools you explicitly allow.
Mast runs on a dedicated cloud VM. No shared filesystem, no noisy neighbors. Memory and state are scoped to your workspace.
Mast authenticates to your tools through OAuth scopes you approve. It never sees API keys, never stores passwords.
Choose exactly which actions each integration exposes. Mast can read Gmail but only send to allowed domains.
It gets sharper every week
Every edit, every approval, every “actually, do it this way”. Mast remembers and applies it next time. No prompts to maintain, no settings to tweak. You work, it learns.
Simple pricing. Pay for what you use.
A flat base keeps Mast running 24/7 on every channel. Credits cover what actually consumes resources: model tokens, voice, image generation.
Per workspace. Cancel anytime.
Start your workspace- Mast on every channel: email, Slack, iOS, web
- Unlimited threads, tasks, and team members
- Persistent memory across every conversation
- All integrations and tool approvals
- Audit log and human-in-the-loop controls
Top up in packs from $10 to $500. Credits never expire.
- Model tokens. Claude, GPT, Grok at-cost pass-through
- Voice. Speech-to-text and synthesis on the iOS app
- Image generation. When Mast creates or edits visuals
- Premium tools. Web search, enrichment, document parsing
Download Mast
On your desk, in your Slack, on your phone. The Mac app sits in your menu bar. The web app runs everywhere else.
Frequently asked questions
Mast is an AI teammate, not a tool. You hire it like you'd hire a coworker. It gets a name, a dedicated inbox, access to your tools, and a memory of every conversation. It runs 24/7 and ships real work: drafted emails, resolved tickets, reconciled books, scheduled calls.
Mast runs on its own dedicated cloud VM with a persistent execution loop. It wakes on a schedule or in response to triggers (new email, calendar event, CRM update), executes its tasks, and goes back to sleep. State persists across every session via the mast memory engine.
Yes. All data is encrypted at rest and in transit. Mast's memory is isolated per workspace. We never train on your data. You can revoke any integration and wipe its memory at any time.
Your pick. Mast works with Anthropic (Claude Sonnet, Opus, Haiku), OpenAI (GPT-5 family and reasoning models), and xAI (Grok-4). It can switch models per task: cheap models for triage, sharp ones for memos. You can also connect your existing ChatGPT subscription so it routes through your Plus/Pro/Team credits.
Mast proposes actions before executing them. You review and approve or reject each one: sending an email, updating a record, scheduling a meeting. Nothing leaves your workspace without your explicit approval.
Yes. Mast is shared across your workspace. Anyone on the team can @-mention it in Slack, email it, or hand it a task. It remembers every conversation so context carries between teammates.
Mast is $50 per month per workspace, plus credits for usage. The base covers unlimited threads, all channels, memory, and integrations. Credits ($1 each, never expire) cover LLM tokens, voice, image generation, and premium tools. Bring your own model keys to skip LLM credit charges entirely.
No. Mast runs on system keys by default and you pay only in credits. If you'd rather, you can connect your own OpenAI, Anthropic, or xAI keys (or your ChatGPT subscription) and skip the LLM credit charges entirely. Switch any time from settings.
Meet your new teammate
$50 a month for the workspace. Credits at $1 each, never expire. Cancel anytime.