Descarty
日本語Download overviewRequest a demo
Request a demo
Menu
Request a 15-minute demoDownload the overview日本語

Run it wherever your company needs it

Kitewell runs on employees' PCs, and just as well in your own Azure, AWS, or Google Cloud, on your servers, or in a closed network. For AI, use your own Azure OpenAI, Amazon Bedrock, Google Vertex AI, or in-house models. Called from your systems, it runs as an unattended part of them.

Discuss a PoCDownload the overview
  • Runs on your PCs, in your cloud, or in a closed network
  • Use your own Azure OpenAI or in-house AI
  • PoC free unless it meets the agreed criteria

Where it runs

Workflows and knowledge are the same wherever they run. You can run the PoC on an employee's PC and production in your own cloud.

Employees' PCs

Run it on each person's Mac or Windows PC. It is the fastest way to start.

  • It runs with that person's sign-ins, so no shared robot account
  • The team shares the same workflows and knowledge

Your cloud

Deploy it to a VM in your own Azure, AWS, or Google Cloud. It keeps running at night and on holidays with nobody signed in.

  • Runs inside your tenant, network, and security policies
  • A fixed outbound IP for partner sites that allow-list addresses
  • People build and fix workflows from a browser

Your servers and closed networks

It runs inside networks with no internet access. With an in-house model, nothing leaves your network.

  • Licensing, team sync, and updates stay inside your environment
  • Fits work on closed factory and branch networks

A VM Descarty hosts

Descarty sets up and runs a VM dedicated to you, so you start without a cloud contract or VM administration of your own.

  • A VM never shared with other customers
  • A fixed outbound IP

Inside your systems and products

Call it from your systems by API or webhook and hand it the screen work. You can also offer it to your customers as part of your product.

  • Runs as an unattended step of your process
  • Offered under your own brand

About OEM and embedding

Which AI it uses

Use the AI your company already contracts. You choose the model for each step, so a step that handles personal data can use an in-house model while the rest use another.

Your Azure OpenAI
Use the Azure OpenAI models in your own tenant. What you send is not used for training.
Amazon Bedrock
Use Claude and other models through Bedrock in your own AWS account.
Google Vertex AI
Use Gemini or Claude through Vertex AI in your own Google Cloud project.
In-house and local models
Use an OpenAI-compatible model server of your own, such as vLLM or Ollama. Prompts never leave your network.
The Anthropic, OpenAI, and Gemini APIs
Use them with your own API keys. The paid APIs do not train on what you send by default.
Replay without AI
Operations it has learned repeat the same way with no AI. AI reads the screen only to learn a step and when the screen changes.

Started by your systems, results sent back

Nobody has to press a button. Your system sends a request, Kitewell works the screens, and the result goes back to your system. A person is asked only when a decision is needed.

What starts a run

  • A webhook or REST API call from your system
  • A schedule, at set times or intervals
  • An incoming email
  • An MCP call from Claude, ChatGPT, or your own AI agents
  • Each row of an Excel or CSV list
  • A person running it by hand

Where results go

  • An HTTP request or signed webhook to your system
  • Entered into core systems and business apps with no API
  • Notifications in Slack, Microsoft Teams, or email
  • Files and Excel workbooks
  • Only decisions waiting for a person's approval

Example setups

Three common setups. In a PoC, we fit the setup to your environment.

Entirely inside your Azure

Kitewell is deployed to a VM in your Azure, and AI is Azure OpenAI in the same tenant. Business data never leaves your Azure.

Requests

  • SchedulesEvery morning at 8, at month end
  • PeopleRun from a browser with a company account

Your Azure

  • KitewellA Windows VM that runs with nobody signed in
  • Azure OpenAIModels in the same tenant
  • Internal systemsScreens and APIs over the virtual network

Results

  • Microsoft TeamsResults and approval requests
  • Partner websitesWorked from a fixed outbound IP

A closed network with an in-house model

It runs inside a network with no internet access, on your own server and your own model. Nothing leaves the network.

Requests

  • SchedulesOvernight batches and more
  • PeopleRun and approve from an internal browser

Your closed network

  • KitewellYour own server
  • In-house model serverOpenAI-compatible, such as vLLM
  • Core systemsScreens and APIs

Results

  • Internal email and chatResults and approval requests
  • Internal file serversResults as Excel or CSV

An unattended part of your systems

Your system sends a request by webhook or API, Kitewell works the screens, and the result goes back. People step in only when a decision is needed.

Requests

  • Your systemRequests by webhook or REST API
  • AI agentsRequests over MCP

Your cloud

  • KitewellA cloud VM that runs with nobody signed in
  • AI modelBedrock, Vertex AI, or Azure OpenAI
  • Screens it worksBusiness systems and partner sites with no API

Results

  • Your systemResults returned over HTTP
  • PeopleApprove only what needs a decision

Business data stays where you choose

Workflows, run logs, results, and credentials are stored on the computer that runs Kitewell. AI receives only the screens and documents an AI step reads. Unless you choose a VM Descarty hosts, Descarty never receives your business data.

Security and data flow

Frequently asked questions

Can it be deployed to our own Azure?
Yes. A VM is set up in your Azure subscription and Kitewell is deployed to it. AI can be Azure OpenAI in the same tenant, so business data stays inside your Azure. AWS and Google Cloud work the same way.
Can it use models on Amazon Bedrock or Google Vertex AI?
Yes. Choose from Amazon Bedrock in your AWS account, Vertex AI in your Google Cloud project, Azure OpenAI, or an OpenAI-compatible model server of your own. You set the model for each step.
Does it work without internet access?
Yes. It runs on your own server with an in-house model, and licensing, team sync, and updates stay inside your environment. Nothing leaves your network.
Can our systems call Kitewell?
Yes. Start a workflow with a webhook or a REST API call and pass it inputs. Results go back to your system over HTTP, or arrive as a signed webhook. Claude, ChatGPT, and your own AI agents call it over MCP.
Does it run at night and on holidays with nobody signed in?
Yes. On a cloud VM or a dedicated PC, it runs on schedule with nobody signed in, right from a restart.
Can we try it on employees' PCs and move to the cloud later?
Yes. Workflows and knowledge stay as they are; only where they run moves to a VM. You can run the PoC on an employee's PC and production in your own cloud.

Plan the setup for your environment

Tell us where it should run, which AI to use, and which systems to connect, and we propose a setup for the PoC.

Discuss a PoCDownload the overview