Browse this section
Getting started

Quickstart

From a new workspace to your first cited answer in about 15 minutes.

This guide takes you from a new workspace to a Knowledge Search Agent answering questions in Slack or Microsoft Teams, with citations from your own documents.

Before you start

  • A BrewMy{Code} workspace. New workspaces are set up personally by our team — request access if you don't have one yet.
  • Admin rights in the source you want to connect first (for example a Google Drive shared folder or a SharePoint site).
  • Admin rights in Slack or Microsoft Teams to install the app.

1. Create a knowledge base

In the web app, go to Knowledge → New knowledge base. Give it a name (ops-eu, support, finance) and choose the region if your plan offers more than one. A knowledge base is the unit of access: users and agents are granted knowledge bases, not individual sources.

2. Connect a source

Click Add source and choose a connector — for a first test, a Google Drive folder or a SharePoint library with 50–500 documents is ideal. Authorise with an account that can see everything you want indexed; permissions are synced along with the content, so users will still only see what they are allowed to see.

The first sync usually takes a few minutes. You can watch progress under Sources → Sync status.

3. Install the Slack or Teams app

Go to Settings → Apps and install for your workspace. Then invite the app to a channel:

/invite @brewmycode

Grant the channel access to the knowledge base you created under Knowledge → ops-eu → Access.

4. Ask your first question

In the channel:

@brewmycode What is our approval threshold for supplier invoices?

The agent replies in a thread with the answer and numbered citations. Click a citation to open the exact passage in the source. If the knowledge base doesn't contain the answer, the agent says so.

5. Try the API

Create an API key under Settings → API keys and run:

curl https://api.brewmycode.com/v1/queries \
  -H "Authorization: Bearer $BMC_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "knowledge_base": "ops-eu",
    "query": "What is our approval threshold for supplier invoices?",
    "as_user": "marja@example.com"
  }'

The response includes answer, citations[] and confidence. See the Queries endpoint.

Next steps