Ask, and the workflow is built
Say what you want in your own words, and the assistant builds a workflow and proposes it.
Nothing is saved until you choose Apply.
AI RPA · Kitewell
The screen changed. The automation fixed itself.
Kitewell is AI RPA: AI does the clicking in web and desktop apps, and people make the decisions. When a screen change stops a robot built by hand, Kitewell's AI relearns the step and carries on. You build and fix workflows by talking to an AI assistant, and the reason for every fix stays in your team's knowledge.
Watch the film (74 s, with sound)Platforms
- macOS
- Windows
- LinuxComing
A script built by hand
Clickx 550, y 524 button#sign-in
Element not found: button#sign-in
Kitewell, on the same changed page
DoClick Sign inRecovered
Make suretext "Invoices"Completed
The interface comes in six languages, including English and Japanese. Chat and steps work in any language your chosen AI model supports. Based on real Kitewell runs. The supplier portal is fictional.
RPA promised to take the clicking off you. Now AI builds it.
Robots were built by hand and fixed by hand whenever a screen changed. In 74 seconds: you describe the job, AI learns it on the first run and finds the new way when a site changes, it stops for your OK, works through an Excel list, and keeps your passwords on your computer.
What is AI RPA? RPA that AI builds and AI fixes
With traditional RPA, people build scenarios in a dedicated designer and fix them by hand whenever a screen changes. With Kitewell, you say what you want and AI builds it, and when something breaks, AI fixes it.
1
AI builds it
Describe the job in your own words, and AI puts the workflow together. No scenario skills or dedicated designer needed.
2
AI fixes it
When a screen changes, AI reads it again and carries on. When a run stops, ask the AI, and it finds the cause and proposes a fix.
3
People stay in control
People approve the decisions that need judgment, and everything AI did or changed is on record.
When your automation stops, can you fix it?
It stops when a screen changes
A small change to a web system stops the robot.
With KitewellWhen a screen changes, AI reads it again and carries on.
Only the builder can fix it
When that person moves on, nobody can.
With KitewellAs you work, the assistant writes down what each workflow is for and why it was fixed.
Nobody keeps track
Robots multiply across departments, unseen.
With KitewellEvery run and change is recorded, and a robot ledger exports to Excel.
Give AI the work, not your passwords
To hand an AI agent work that needs a sign-in, you usually tell it the password in the chat or let it use the browser you are signed in to. A password told in the chat goes to the AI provider and stays in the history; a signed-in browser puts every account signed in there within the AI's reach.
What the AI gets
Enter %password% in the password field and sign in
Only the password's name
What Kitewell types
The real password, encrypted on your machine
Never in the chat or the logs
The assistant never asks for a password in the chat; it shows a masked field there for you to store it. A value that shows up in a run log is masked, and team sync carries only the name.
Never your everyday browser
Web steps run in a browser kept for workflows, separate from the one you use every day, so the AI cannot reach the mail and admin pages you are signed in to there. You can also limit which sites it may open.
No one stays up to type it
Kitewell types the password, so scheduled runs at night or on weekends need no one standing by to sign in.
What AI can operate
Everything RPA could not reach, in the same workflow.
Web systems
Supplier portals, SaaS, and internal web systems. Write the steps in words, and AI operates Chrome the way a person does. A small screen change does not stop it.
Sign-ins are kept only on that machine.
Desktop applications
Sales and accounting software and dedicated clients with no API. On macOS and Windows, AI reads the screen and works the mouse and keyboard.
It waits while a person is using the mouse or keyboard, and an approval can sit before a save.
Email
Gmail, Microsoft 365, and any IMAP mailbox. Find what arrived, have AI read it, record it in a ticket or ledger, and organize what was handled.
Connections and credentials stay on the machine. A failed email is handled again next run.
APIs, servers, and scripts
Imported OpenAPI operations, servers over SSH, commands and scripts, Docker. They sit in the same workflow as the screen operations.
Credentials are encrypted secrets, never written into the workflow.
Build and fix in the language of the work.
Screen operations are sentences
Write steps such as “enter the login ID”, and AI reads the screen to carry them out, in a browser or a desktop application. A small screen change does not mean rewriting the steps.
Steps are plain sentences, so anyone can check what a workflow will do before it runs.
When it stops, fix it on the spot
From the step that stopped, ask the assistant about the cause. Once fixed, run again from that step.
Completed steps keep their results.
What you teach becomes the team's knowledge
Tell the assistant a fact, a rule, or a fix, and it writes it into the knowledge, then reads it before it answers, whether you build a new workflow or fix one. Whoever asks starts from the same place, so the more you use it, the less anyone has to explain.
The lines it changed are marked and can be undone at once. After reading a web page or a run's output, it asks before saving.
Decide what AI does and what people check
Approvals and human tasks
Add an approval to any step. What was decided, and when, is recorded.
Excel lists, as they are
Open your workbook and say what should happen for each row. Results are written back into each row of your own workbook, and can be undone. Values AI reads come with the quote they came from.
A record of runs and changes
Logs and screens for every run, and every version of every workflow. A robot ledger exports to Excel.
Run on each PC. Share every fix.
The whole team works from the same workflows and knowledge, with no management server to build.
Runs with each person's sign-ins
Workflows run on each person's PC or server with their own sign-ins and credentials. No shared robot account is needed.
Every fix reaches every device
A workflow or page someone fixes reaches the team's devices on its own. On the paid plans (Personal and Team), even an unattended device picks up the new version about every 5 minutes.
Roles for each project
Give each person a role on each project. Each run is recorded on the machine that ran it.
Fixes are shared only within your team. Descarty never uses your workflows or knowledge to train AI, and never shares them with other customers.
Your business data stays in your environment
- Workflows, knowledge, run logs, and credentials stay on your machine
- You choose the AI model, including in-house and local models, and switching keeps your knowledge
- Credentials are encrypted, and AI never sees their values
- Descarty never uses your data to train AI
Open-source Dagu underneath
Kitewell runs on Dagu, the open-source workflow engine Descarty develops. You are never locked in to one vendor.
- GitHub stars
- 4,300+
- License
- GPL-3.0
For the work of your industry and department
Browse by industry
Real estate
Finance & insurance
Leasing
Manufacturing
Logistics
Retail & e-commerce
Insurance agencies
Wholesale & consumer goods
Freight forwarding & customs
Construction
Staffing & recruitment
Enter an Excel list into a web system
For each row of an Excel list of orders, AI enters it into a web system and writes the receipt number back into the row. Only failed rows are redone.
Collect orders and due dates from supplier portals
Every morning, collect new orders and delivery date changes from several customers' supplier portals and, after review, record them in the sales system.
Read invoices into the accounting system
AI reads incoming invoices, checks them against purchase orders, and records them in the accounting system. Invoices whose amounts differ wait for approval.
Check property listings across portals
Your Excel property register is compared with each listing portal, and any difference in rent or availability goes into the row and to the person responsible.
Research a list of prospects
For an Excel list of hundreds of companies, each one's business, locations, and size are collected from its website into its row, with a priority for sales.
Screen new business partners from public sources
For each new partner on an Excel list, AI searches news and public sources and flags anything to check. A person decides, and the result stays in the row.
Everything AI RPA needs
- Desktop applications on macOS and Windows
- Read Excel workbooks and write results back
- Knowledge the whole team shares
- Robot ledger export to Excel
- Team project sync with roles
- Find, organize, and send email (Gmail, Microsoft 365, IMAP)
- Schedules and catch-up of missed runs
- Branches and loops
- API import (OpenAPI)
- Server operations over SSH
- File transfer
- Docker containers
- AI agents such as Claude Code
- Alerts to Slack, Teams, and email
- Queues to limit concurrent runs
- Backup and restore
- MCP and REST API
- Interface in six languages
Start by testing it on your own work
Choose one to three real tasks, and a PoC measures the effect on the same screens and data as production.
1
Interview
We learn about your work, problems, and RPA.
2
Selection
Agree on the tasks and success criteria.
3
Build
Build workflows on your real screens and data.
4
Verify
Run them on real data and report the results.
Frequently asked questions
- What is AI RPA?
- RPA that AI builds and AI fixes. Describe the job in your own words, and AI builds a workflow that reads the screen and works web and desktop apps the way a person does. Traditional RPA tends to stop when a screen changes; with AI RPA, AI reads the changed screen and carries on.
- How is it different from traditional RPA?
- You ask AI to build it and to fix it. There is no scenario to assemble in a dedicated designer, and when a screen changes, AI fixes it. The reason for every fix stays in the team's knowledge, so automation keeps running when people change. Work that needs judgment, such as reading invoices, can be automated too, by pairing what AI reads with a person's approval.
- How is it different from asking an AI agent?
- An AI agent is good at doing a task there and then, for the person beside it. Kitewell is AI RPA for running business work every day, even unattended. A workflow AI builds is saved only after a person reviews it, it waits for approval where judgment is needed, and every run is recorded. When the work needs a sign-in, the AI never gets the password: Kitewell types it.
- What does the AI cost to run?
- AI is used to handle new operations and changed screens, and to read screens and documents, not for every click on every run, which keeps AI fees down. You choose the AI model, including in-house and local models.
- Can Kitewell replace the RPA robots we run today?
- Yes. Start with the robots that stop most often and matter most when they do. A robot that repeats fixed operations at high volume on an internal system whose screens rarely change can stay where it is. You can try one robot in a two-week PoC.
- Does our business data leave our environment?
- Workflows, knowledge, run logs, and credentials stay on your machines. What the AI reads from a screen goes to the AI model you choose; with an in-house or local model, nothing leaves your network, though small local models usually cannot operate screens. Descarty never uses your data to train AI.
- Which operating systems does it run on?
- macOS and Windows, with Linux coming. AI works websites in Chrome, and desktop apps with the mouse and keyboard, the way a person does.
From automation that breaks to automation you can fix
Tell us which work you have in mind, and we will propose how a PoC would run.
For your own work, start free from the product site. Go to the Kitewell site