Assista — Your AI team is already working

Your AI team is already working.

Type one goal. A CEO picks the specialists, connects your apps, and ships drafts to your inbox — usually before lunch tomorrow. It all runs on your own Claude, so we never mark up or meter your AI.

Bring your own Claude · unmetered AI · cancel in two clicks

Live · your team, in motion

Last action · 4s ago

Goals in Progress

The teammates

One job each. Done well.

Sales · Inbound

Qualify every lead. In your voice.

An SDR teammate reads inbound, enriches the contact, and drafts a reply in your tone — so by the time you read the thread, the easy decision is one click away.

Visitor info
Company: Nova Commerce
ARR: $6M
Team: 28 employees
Offer: Multi-product
Goal: reduce sales cycle

Support · Triage

Triage every ticket. Draft every reply.

A support teammate reads every inbound, classifies it, drafts a reply from your help docs and past tickets, and pings you only for the messy ones.

Drafted replies: 14/14
Hi Sara — totally hear you. The invoice export now lives under Billing → Exports. I've attached your last three. Let me know if…

14 ready for your review

Product · Onboarding

Walk every new user to their first win.

A guide teammate knows your product cold, navigates inside your UI alongside the user, and helps them reach their first win — without a support ticket touching your inbox.

Hello! I'm an onboarding agent. How can I help?

Marketing · Content

Post on cadence. In your voice. Without you.

A writer teammate studies your tone from what you've already shipped, drafts LinkedIn threads, X posts, and newsletter sections on a weekly cadence, and queues each one for your one-tap approval.

LinkedIn · draft queued

Engineering · Triage

Triage every issue. Unblock every PR.

An engineering teammate watches GitHub and Linear, labels new issues into your taxonomy, drafts the first investigation comment, and pings the right human when work stalls.

triaged

Ops · Weekly status

Monday morning, written for you.

An ops teammate pulls from GitHub, Linear, HubSpot, Stripe, and Slack, then writes a one-page digest in your team's voice. Standup becomes a read, not a rebuild.

#general Mon 9:00

Week of May 19

What the team ships

Real work. Not a chatbot reply.

These are the shapes of work your teammates produce — emails, posts, briefs, pitches. Drafted, never auto-sent.

Connections

Works inside your stack. Not next to it.

Teammates read, write, search, summarize, and trigger actions inside the apps your company already uses. Each gets a scoped OAuth token — least privilege per role, revocable from settings.

Gmail Slack Notion LinkedIn GitHub Linear Jira Asana HubSpot Salesforce Intercom Zendesk Stripe Shopify Dropbox Drive Zoom Calendly Figma X Reddit Vercel Airtable Mailchimp

Pricing

Hire your first teammate today.

You pay for the teammate — the schedule, the apps, the memory, the follow-through. The AI runs on your own Claude, unmetered, with no markup from us. Cancel in two clicks.

Starter

$20/month
For solo work and small projects.

Most Popular

Pro
$35/month
For founders running the whole team.

Team

$40/seat/month
Shared teammates, one workspace.

Questions

The honest answers.

What actually happens after I set a goal?
A CEO reads the goal, breaks it into workstreams, and picks the specialists you'd hire — CMO, SDR, support, CTO, ops, finance. Each spins up with the right tools and scopes, then starts working on a Kanban board you can inspect at any moment.

How is this different from a chatbot or AI workflow builder?
A chatbot answers questions. A workflow builder runs a flow you designed. Assista is an operating team — it decides what work needs to happen, assigns it to specialist teammates, executes across your apps, and wakes up on triggers and schedules to keep moving. You stay in the loop on what matters.

Which apps do the teammates work inside?
1,000+ apps — Gmail, Slack, GitHub, Stripe, Calendar, Notion, Linear, HubSpot, LinkedIn, Reddit, Drive, and more. Connect once with OAuth. Each teammate gets a scoped token, least privilege per role.

Do teammates act on their own, or do I approve everything?
Autonomous where useful, controlled where it matters. Teammates draft, prepare, research, update, and suggest on their own. For sensitive actions — sending external messages, changing shared state, posting publicly — they wait for your approval before anything goes out.

Can teammates react to events and run on a schedule?
Yes. A Stripe payment fails, a GitHub issue opens, a Slack mention hits a blocker, a Reddit post mentions your product — Assista routes each to the right teammate and continues the work. You can also schedule recurring goals like a weekly ops digest.

Does Assista remember what it learns about my company?
Yes. Every task teaches the system more about your product, tone, ICP, customers, objections, decisions, and recurring workflows. Your teammates don't restart from zero — every run starts smarter.

Why do I connect my own Claude?
Your teammates think and write with Claude, and we'd rather you pay Anthropic directly for that — at their rates, with zero markup or metering from us — than have us resell you tokens and show you an 'out of credits' wall. It also means your work runs on your own AI account, which you control and can revoke anytime. You pay us for the employee: the schedule, the apps, the memory, and the follow-through.

What does the AI cost me?
You pay Anthropic directly for what your teammates use, at their rates. A busy teammate typically runs a few dollars of Claude a month. If you already pay for Claude Pro or Max, your subscription usually covers it — so your extra AI cost can be $0.

I don't have Claude yet — is that a problem?
Not at all, and it's free to start: Anthropic gives new accounts $5 of free credit — no card needed, just a phone number to verify — which is plenty to try your team out. During setup we walk you through it in about two minutes, all in plain English. Or, if you already pay for Claude, just connect the plan you have.

What models can my teammates use?
Any Claude model — Haiku, Sonnet, or Opus — on every plan. Because the AI runs on your own account, nothing is gated behind a tier. Your teammates pick the right model for each task, and you can steer it.

Do I still need a paid plan if I bring my own Claude?
Drafting and reviewing your teammates' work is free. A paid plan turns on the parts that make it an employee — running on a schedule, publishing and sending, unlimited teammates. Your Claude covers the thinking; your plan covers the showing-up.

Is my Claude key safe?
Yes. It's encrypted at rest, used only to run your own workspace's work, and revocable by you at console.anthropic.com or in your Claude account at any moment. We never see your teammates' AI usage billed to us, because it isn't.

Does Assista train models on my data?
No. Your data, your tools, your work — none of it goes back into training. Period.