MILabs
Tools

Chatbot UI self-hosted AI chat for business teams

How Chatbot UI, an open-source self-hosted AI chat interface, fits small business use, and how MILabs sets it up.

Updated 2026-09-04

What Chatbot UI is

Chatbot UI is an open-source, self-hosted AI chat interface created by McKay Wrigley. It gives people a familiar, chat-style screen for typing questions and reading replies, and its code is published on GitHub, so a team can inspect it, change it and host it on its own server or cloud account. Each message goes to a language model through an API key the business provides, and the answer appears in the same window. The interface does not decide what the model knows. That comes from the chosen model and from any material the team deliberately adds. Teams tend to look at Chatbot UI when they want one shared chat screen for staff, or more control over where the software runs than a public website offers. It handles the conversation itself and does not run multi-step workflows on its own. Its fit depends on the business having someone willing to keep hosting, updates and keys in good order.

McKay Wrigley: official site

Chatbot UI in plain language

A chat interface is the part of an AI product where a person types a question and reads the reply. Most people meet one through a public website run by a large provider. A second kind is software a business installs itself and points at a model provider through a key the business holds. The chat looks much the same in both cases. What differs is where the software runs, who maintains it and whose account pays for each message.

Chatbot UI belongs to the self-hosted kind. Its source code is public, so a technically capable person can read it, adapt the screens and decide how it is deployed. Maintenance then sits with whoever runs the copy. Security patches and dependency updates arrive from outside, and somebody has to apply them. Treating the installation as finished means falling behind over time.

Cost and control follow the same split. A hosted assistant bundles model, screen and upkeep into one service, so the bill is easier to predict, though choices stay within the vendor's offer. A self-hosted screen separates those pieces. The business pays its model provider for usage and pays for the server that runs the software, and it can switch models later while staff keep the same screen.

What businesses use it for

  • Shared chat windows where staff draft emails, summaries and first versions of documents
  • Questions put to a model with material the team chooses to paste in and may share
  • Conversations kept on a server the business controls, useful where privacy matters
  • Draft replies to customer enquiries, reviewed by a person before anything is sent
  • A simple internal assistant for quick answers during a busy stretch

Who it suits

  • Small teams wanting a shared, private chat screen for everyday writing and research
  • Businesses with a technical contact who can keep a self-hosted app updated and secure
  • Organisations that want more control over privacy than a public chat site offers

How MILabs sets it up

Setup starts with a conversation with the owner or team lead, through the enquiry form. We agree who will use the tool, which model provider the business already holds an account with, and whether the software should run on a server the business owns or a cloud account it manages. We then connect the API key, configure hosting and check that each person's access matches their role. Where the business already uses a document store, shared drive or messaging app, we look at what could be linked and what should stay separate. The team agrees in writing who looks after updates, keys and backups, and we write a short staff guide. The business keeps ownership of its keys and data throughout.

Builds it fits

Industries it fits

Chatbot UI in practice

The day in a business running Chatbot UI usually starts with a few people who already know what they want from it. A bookkeeper might paste in a list of unusual transactions and ask for a plain-English explanation to share with a client. A paralegal might request the structure of a letter and fill in the facts afterwards. The useful habit is to keep each question specific and to label any pasted material so the answer can be checked against it.

The shape of the work depends on the boundaries the team sets. Most businesses settle on two lists: material that may go into the tool, such as published policies and templates, and material that stays out, such as identifiable client files. Someone reviews both as new tasks appear. Outputs are treated as drafts, and a person checks figures and names before anything leaves the office.

Administration is quieter but steady. One person keeps the model keys and access list current, watches for notices from the project and the model provider, and tests a restore at an agreed interval. Staff who see an odd answer report it to that person rather than changing the setup themselves. Over time the team gathers prompts that suit its own material.

good questions, straight answers

Frequently asked questions

What do businesses use Chatbot UI for?
Most teams treat it as a shared chat window for drafting, summarising and research, with room to paste in documents they are allowed to share. Because it can run on infrastructure the business controls, it also suits conversations involving client or staff information the owner wants kept in-house. It is a front end, so answer quality depends on the model behind it and the material the team supplies.
Does Chatbot UI need technical skills to run?
Some technical ability is needed, since any self-hosted application involves a server or cloud account, model API keys, updates and backups. Many small teams would rather not run those pieces themselves. MILabs can handle the setup and agree with the owner who looks after it afterwards, so day-to-day staff only need to log in and type their questions. Their job is simply to ask good questions.
Can Chatbot UI connect to the software I already use?
Yes, but the link is usually built rather than switched on. Chatbot UI is a chat interface, so it does not read accounting or booking software by default. Where a connection is worth making, we look at the systems the business already uses and decide what should pass through and what should stay separate, then build and test that link before anyone relies on it.
How does Chatbot UI compare with other tools for the same job?
Hosted assistants such as ChatGPT, Claude or Microsoft 365 Copilot are simpler to start with and need less upkeep. Open-source options like Open WebUI or LibreChat offer similar self-hosted chat screens, and choosing between them mostly comes down to features and how the team wants to maintain the software. The right fit depends on who will use it, where data should live and who will look after it.
How do I get my team to use Chatbot UI?
Start with a small group who already have a real task, such as drafting replies or summarising notes. Share a short written guide with examples of useful prompts and clear rules about what can be pasted in. Name one person to answer questions, and review how the tool is being used with the team regularly so the guidance can improve.
let's make something happen

Workflow automation: how we can help

Choose how you want to work on this with MILabs. Both options start with an enquiry.

Discovery is free, no obligation from the first chat.

Let us build it for you

We build it and look after it.

Build it yourself, we'll help

You build it with our direction, teaching and a weekly check-in.