# One by Lughlabs: AI teammates with their own computers

> One by Lughlabs is a platform for AI teammates that work on their own computers. An Architect maps how your business runs and designs a crew of specialized bots; you approve it before it’s built. Each bot works in the apps its job needs, from a catalogue of about 1,500, remembers what it learns, runs routines on a schedule and hands work to the next. Voice agents answer your phone line, and decisions that need you reach your phone. Bring your own model keys or use included credits. Creator is $40 a month or $240 a year; Business is $100 a month; Business Pro, on semi-private hardware, $400 or $1,000 as Business Max; Enterprise, a private server run for you, from $4,000. Lughlabs makes One.

URL: https://lughlabs.ai/

## One handles every order while you cook with your family.

[Try One free](https://one.lughlabs.ai/sign-up)

- 500 free credits
- No card needed

How it works

### One maps your entire business.

#### 01 Start with what you have.

Like a consultant, your Architect maps how your business runs and what matters most, then designs the crew to grow it. You approve it before it’s built.

#### 02 Bring or build your tools.

About 1,500 apps, connected once for the whole crew: some in one click, others with your own key. Missing one? Build your own with MCP or OpenAPI.

#### 03 Meet each bot at its desk.

Each one works on its own computer, in the apps its job needs. It tracks its numbers, runs its routines and passes work to the next. Correct it once and it remembers.

#### 04 Evolve with your Architect.

Your Architect sees every number and handoff across the crew, and keeps what the business learns. Each week it suggests what to improve, with its reasons, and changes only what you ask for or approve.

#### 05 Step away, stay in charge.

Routines keep their schedule and calls get answered. What needs your decision reaches your phone, and you settle it there. Back at your desk, everything the crew did is in Activity.

Product names and logos are trademarks of their respective owners. Their use does not imply endorsement.

Who it’s for

### What you’d hand off first.

Your business SaaS startup

#### Tickets answered before you’re in, the two that need you waiting.

Kai answers support from your help docs, Lena follows up demo requests, Ada answers security reviews and Nell drafts the investor update.

Your business SaaS startup Interactive example

Today

The crew takes first

- Answer this week’s tickets from the help docs
- Follow up everyone who filled in the demo form
- Answer a 60-question security review from your policies
- Draft this month’s investor update from the close

People

- **You** Founder

Crew

- **Kai** Customer support
- **Lena** Sales follow-ups
- **Ada** Security reviews
- **Nell** Books and investor updates

Works in

- Intercom
- HubSpot
- Gmail
- Google Docs
- Stripe

[One for Teams](https://lughlabs.ai/one/teams) [One for Individuals](https://lughlabs.ai/one/individuals)

An example. Every person and bot named here is fictional.

Make it work for you

### Start from the job.

- [Daily Operations](https://lughlabs.ai/use-cases/operations)
- [Compliance](https://lughlabs.ai/use-cases/compliance)
- [Accounting](https://lughlabs.ai/use-cases/accounting)
- [Content Creation](https://lughlabs.ai/use-cases/content-creation)
- [Recruiting](https://lughlabs.ai/use-cases/recruiting)
- [Sales & Marketing Campaigns](https://lughlabs.ai/use-cases/sales-and-marketing)

Features

### See what your crew can do.

Northline Studio is a fictional company for demo purposes. Every stage runs One’s own interface.

Inside One [Chapter 1: One mind](https://lughlabs.ai/#intelligence) [Chapter 2: Voice](https://lughlabs.ai/#voice) [Chapter 3: Computer](https://lughlabs.ai/#computer) [Chapter 4: Routines](https://lughlabs.ai/#routines) [Chapter 5: Self-evolve](https://lughlabs.ai/#evolve) [Chapter 6: Build](https://lughlabs.ai/#build) [Chapter 7: Made for you](https://lughlabs.ai/#custom)

#### Your bots ask each other, share one Wiki and follow one Vibeboard. Brief one, and the crew moves.

Nora asks Maya whether the follow-up email can promise Saturday delivery. Maya asks Alex and Sage at once, Alex checks the studio’s shop, and Maya answers once, showing who helped and what they said. A café owner tells Sage on a call that cafés order the Ridge bowl in sets of 12; Sage proposes it to the Wiki and Nora accepts it for the whole crew. On Vibeboard, wholesale enquiries rise from 3 to 11 in a month; Nora asks the Architect to get the crew on it, it proposes a task for each bot, and once she applies them each bot takes its part.

#### Describe the phone agent you want. Your bot builds it, answers your calls and hands the work to the crew.

Nora asks Sage for a wholesale line. Sage builds a phone agent in the Voice app: its steps in the window, the call flow drawn node by node from greeting to wrap-up, a test with callers that passes, then Test Agent beside the flow, and asks Nora to publish it to its number. Ana, who runs a café, calls the line: on her phone she asks for 24 Ridge bowls; inside One the live call shows her words and Sage’s, each step of the flow, the stock check and Nora’s calendar, where Sage books a call back for Tuesday at 10:30, and a text reaches Ana’s phone. When the call ends, it becomes a row in Sage’s Enquiries and Sage hands the work to Maya and Alex. Maya’s quote is a calculator you can try: sets of 12 times the price per set, with the total including VAT; sending it emails Ana, and Alex holds the bowls.

[Read the voice and phone guide](https://lughlabs.ai/docs/voice)

#### Every bot has its own desktop, browser, terminal and files. It signs in and uses your apps the way you would.

Inside Alex’s own computer, full screen, with his cursor at work. He signs in to Stockyard, a wholesale marketplace, with a login Nora saved in One’s encrypted vault; Stockyard sends a sign-in request to Nora’s phone, she approves, and Alex’s browser stays signed in there, as it is to the studio’s shop, its mail and Kiln Supply. Asked to make the Ridge bowl €58 everywhere, he changes it in a spreadsheet, the shop and Stockyard, and checks all three in the terminal. Ordering glaze at Kiln Supply, he stops at the card field and asks for you; you take control, enter the card, hand back, and he places the order and saves the invoice to the shared folder. Then you show him how to export this week’s orders; he writes the steps down as a draft skill and repeats them in a test run.

#### Tell a bot what to do and when. It runs on a schedule, or the moment something happens.

Nora writes one sentence to Alex: every weekday at 7, send me yesterday’s orders and anything unusual. Alex turns it into a routine in his panel: its name, its instruction using the Weekly orders export skill, the schedule Weekdays at 7:00 AM, and a notification when it posts. The clock moves to Friday 7:00, the routine runs by itself, and Nora’s phone shows the result: 31 orders, €2,140, and six Ridge bowls to one café in Leeds. The next Tuesday, that café, Juniper Café, sends the wholesale form on the studio’s website; the form starts Sage’s routine at once. Sage reads it, looks the café up, and asks to call back, because a run started by a webhook asks before a call; Nora allows it, the café’s phone rings, and Sage writes a record and hands the price list to Maya, three minutes after the form. Then the crew’s week, as a lane for each bot from Monday to Monday: every scheduled run, the form, an email, and Ridge bowl stock falling under 10, each lighting up as the week passes; you can move through the week with a slider. On Monday at 8:00 Maya’s weekly brief asks Alex and Sage for their parts and sends one file to Nora’s phone. Last, Alex’s runs this week, as rows of his Routine runs table in Records: what started each, how long it took and what it made; Tuesday’s Stockyard sync failed because the sign-in had expired, and Alex marked it recovered after Nora approved signing in again; Nora opens Wednesday’s Morning orders record.

[See routines in operations](https://lughlabs.ai/use-cases/operations)

#### Each bot tracks your outcomes and tests what moves them. Your Architect turns what they learn into your next strategy.

Twelve weeks into Northline’s fictional One. Maya’s Overview shows the numbers she is judged on, each with its source, her twelve weeks, and her reading: subject lines that name a piece seem to get more opens. She tests it on under 10% of the list: two subject lines, two inboxes, 41% against 29%. She sends the winner to everyone, keeps the lesson in her memory, and the list’s line bends up. On its weekly look after the space, the Architect reads on Vibeboard that wholesale has tripled in four weeks while retail is level, and proposes the Strategy board’s next version (lead with wholesale, new work for Maya, Alex and Sage, a target), a routine for Sage and a new bot, Hollis. Each waits for Nora: she applies the board and the routine, and Hollis waits. History lists twelve weeks of the Architect’s changes, each with who decided it.

#### Bots write the tools they need. Your engineer bot builds your sites and apps, publishes them and keeps them running.

Alex has compared three kiln suppliers’ price lists by hand three weeks running, so he writes a script in the Terminal on his own computer, installs what it needs, tests it against last week’s lists and saves it as a skill, Supplier check, which joins his skills under Advanced › Knowledge; he uses it at once, and typing / in the composer lists it with the space’s other skills. Remy, the crew’s new engineer bot, builds a wholesale portal from one message: a plan (pages, trade prices, sets of 12 from the Wiki, orders by email, its address), its files in Library, its build and tests in Terminal, then code on its own computer and a preview that assembles beside it; Nora comments on the preview’s button and Remy changes it. Remy asks before publishing, and Nora allows the first release to its address, northline-wholesale.lughlabs.app; outside One, Juniper Café orders two sets of 12 Ridge bowls, the order arrives as an email, and Remy’s routine writes it into Orders, with handoffs to Maya to confirm it and Alex to hold the stock, and a skill that lets the crew look orders up. When Nora changes the Low carafe’s trade price in the Wiki, Remy’s afternoon check finds the portal still showing the old one, changes the price with a test, checks it again and asks Nora before publishing release 2; she allows it, and the portal is right again, with release 1 kept to put back.

#### Ask for what you need. We build it in the open for everyone, or just for you.

A drawn workshop with a glass front on the street, where the community stands. One of them sends a request: plans for their builds, shown by a screenshot of FreeCAD with a building modelled in 3D. It lands on the workshop's board under two others, a sales map (QGIS) and 3D renders (a clip of Blender rendering), and the street votes it to the top. The crew takes it on and builds it while everyone watches: the timeline reads Asked, Voted, Building, Shipped. A verified developer comes in off the street, lanyard on, and is handed the AI credits to build with. It ships the same day, and a tile for it goes up on the shelf of everything shipped. Then the atelier, for Teams Max and Enterprise: a client comes in through a private door, and Dapper the tailor takes its measure while the scope and then the plan are agreed. Its request, its own circuit boards, shown by a screenshot of KiCad, is built for it alone, sealed in a crate and carried to its own cabinet. The requests are illustrations; the products are real.

The requests are illustrations. The products are real, and belong to their owners; their use does not imply endorsement. [Screenshot credits](https://lughlabs.ai/licenses#screenshots)

[Ask for a custom solution](https://lughlabs.ai/contact?interest=Custom+solution) [Create your One account](https://one.lughlabs.ai/sign-up) [One on a private server](https://lughlabs.ai/one/enterprise)

**500 free credits**No card needed

[Try One](https://one.lughlabs.ai/sign-up)