MILabs
Tools

Big-AGI self-hosted AI chat for small business teams

Big-AGI is a self-hosted AI chat interface. See what it suits and how MILabs sets it up for a business.

Updated 2026-07-26

What Big-AGI is

Big-AGI is a chat interface for working with AI models, built by Enrico Ros. It is designed to be self-hosted, so the software runs on a server the business chooses rather than only inside a vendor's hosted service. Staff use it through a web browser to send questions and documents to the language models connected to it, and conversations stay together in one place. For a small business, the main point is control. The business decides where the server runs and which people are given access. Big-AGI is a front end rather than a complete business system, so it does not replace a CRM, accounting package or helpdesk. Sessions with the owner or team lead run by video call, and MILabs is a Sydney agency that works remotely. MILabs sets up Big-AGI when a client wants a shared AI workspace hosted on its own server rather than in a general public chat app. The work covers which model providers are connected, how the server is hosted and who can reach it.

Enrico Ros: official site

Big-AGI in plain language

A chat front end is the window people type into when they talk to an AI model. The model does the generating, usually run by a company such as OpenAI, Anthropic or Google, while the front end handles the conversation, the list of past chats and the settings. Many people know only the front end of a hosted service, but a separate one can sit between staff and several models at once.

Self-hosting changes where that layer lives. The software is installed on a machine the business controls, whether that is a rented cloud server or a computer on the premises, and it is configured to call the model providers the business has agreed to use. Requests travel to those providers over the internet unless the models are also run locally, so self-hosting mainly concerns control of the interface, the stored history and the access rules. It does not keep every request in-house.

The trade is maintenance. Someone has to install updates, rotate keys, check that the server is secure and decide who can log in. A business that enjoys those tasks gets more say over its setup. A business that does not may find a hosted product less work for much the same chat experience.

What businesses use it for

  • Drafting emails, proposals and product listings that staff then review
  • Answering internal questions about policies and procedures from supplied documents
  • Comparing how different connected models answer the same question
  • Summarising long meeting notes or reports into a list of actions
  • Keeping research sessions in one place so a team can return to them

Who it suits

  • Owners who want staff to use AI on a server the business controls
  • Teams with an IT contact who can keep the software current
  • Businesses that need a shared drafting and research space rather than a finished system

How MILabs sets it up

MILabs starts with the owner or team lead to agree the hosting choice the business needs, including where the server will run and which model providers it will connect to. Next, the team links the tools and document stores the business already has, where the setup allows it, so that only agreed material reaches the models. Access is then limited to named staff, with permissions written down before the tool opens to anyone. Finally, the owner and MILabs agree who looks after updates, keys and day-to-day questions. A short written note records those decisions, so the business keeps a clear reference after handover. Sessions run by video call, and MILabs is a Sydney agency that works remotely.

Builds it fits

Industries it fits

Big-AGI in practice

Inside a business that runs a self-hosted chat front end, the daily pattern looks like a shared drafting space. A marketing coordinator pastes a rough brief and asks for several headline options, then edits the strongest before it goes anywhere. A bookkeeper asks for a tax rule explained in plain words, then checks the answer against the official source before a client sees it. Each person works in their own conversations, and the owner sets how long the team keeps them.

The shape of the work depends on the rules set at the start. Access tends to go to named staff rather than the whole company, and a short written guide tells people what material is appropriate to paste in. Where a team relies on a policy manual or product sheet, someone prepares the current version so the tool answers from it rather than from an old copy.

Most of the effort goes into habit rather than configuration. Staff learn to state the audience, length and purpose of a request, and to treat every figure and citation as something to verify. The owner or team lead usually keeps a routine check of accounts, access and model settings, so the setup stays matched to the business it was built for.

good questions, straight answers

Frequently asked questions

When is Big-AGI the wrong choice?
Big-AGI fits a team that wants a shared chat space on a server it controls. A business that needs a finished system for records, invoices or customer tickets will be better served by a hosted product or a workflow tool, because Big-AGI is a chat front end and does not run those processes. A team with someone who can look after a server and keep software current gets the most from it.
Can my team use Big-AGI without training?
Yes, staff who have used a chat window before can usually start after a short walkthrough from the owner or team lead, because the basic task is typing a question and reading the reply. Strong results take a little more experience, especially when writing clear requests and checking answers against source material. The handover note from MILabs records the team's agreed habits so new staff have a reference to follow.
Can MILabs build on top of Big-AGI?
Yes, MILabs can build around Big-AGI, for example by connecting it to a document store the business already holds, or by writing a small layer that prepares prompts for common tasks. Each build is shaped by what the self-hosted installation supports, so MILabs confirms that during setup. Changes are described in writing so the business can see what has been added and who maintains it.
What does a Big-AGI setup need from me?
Start by deciding who will use the tool, which model providers the business is happy to connect, and where the server should run. Name one person who will handle updates, access changes and questions after handover. Gathering the documents staff should use as reference material helps too, because the setup can then be tested against real work. Sessions run by video call.
What does Big-AGI not do?
Big-AGI gives people a place to talk with AI models, and it does not manage customer records, send invoices or run scheduled workflows on its own. It also leaves the check of each answer to people, so staff review drafts and figures before they are used. Any link to other business systems is separate work, and it is scoped at the start so the team knows what it covers.
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.