Using Boomi Companion for Data Integration with Claude
This topic walks you through operating Boomi Data Integration (BDI) using Boomi Companion, from connecting your account to running your first request in plain language. It uses the bc-bdi plugin, which packages the boomi-bdi skill for Claude. BDI moves data into a warehouse or lake, transforms it there, and can send curated data back out to other tools.
You can ask in plain language, and Boomi Companion reads from or acts on your BDI account and shows you the result. BDI adds one difference worth noting up front: reading data runs freely, but any request that changes something waits for your approval first.
Install the bc-bdi plugin
-
In Claude Code, run:
/plugin marketplace add OfficialBoomi/boomi-companion
/plugin install bc-bdi@boomi-companionFor more information, refer to Installing the plugin on Claude Code.
-
For Claude Cowork, refer to Installing the plugin on Claude Cowork.
Prerequisites
- Access to a Boomi Data Integration account and an API token for that account.
- A connection to your source or target system already configured in the BDI console.
- If you don't know your
environment id, omit that line entirely - an emptyBDI_ENVIRONMENT_ID=line blocks the other commands. Ask Boomi Companion to list your environments, then add it.
Core workflow
Step 1: Connect your account
Boomi Companion reads credentials from the .env file in your workspace directory - the folder you launch Claude Code from. There is one .env per workspace, shared by all Boomi Companion plugins.
If you already have a .env file for another plugin, add the four variables below to it; the BDI_ keys won't conflict. BDI has its own console, API, and tokens, so nothing carries over from Boomi Integration - you'll still need to generate a new BDI token.
If this is your first Boomi Companion plugin, create the .env file yourself in that folder. Boomi Companion won't create or edit it: the command-line tools read your token straight from the file when they call the BDI API, so it never enters your conversation.
BDI_API_URL=https://api.rivery.io
BDI_API_TOKEN=<your token>
BDI_ACCOUNT_ID=<account id>
BDI_ENVIRONMENT_ID=<environment id>
Generate the token in the BDI console, under My Profile, then API Tokens. You can find both your account id and environment id in your BDI console URL, which follows the pattern https://console.rivery.io/dashboard/<account-id>/<environment-id>/....
Select the host that matches your region and use the full value, including https://, as your BDI_API_URL:
| Region | Host |
|---|---|
| US | https://api.rivery.io |
| EU | https://api.eu-west-1.rivery.io |
| IL | https://api.il-central-1.rivery.io |
| AU | https://api.ap-southeast-2.rivery.io |
Step 2: Ask Boomi Companion in plain language
Boomi Companion activates on its own whenever you mention BDI, data flows, extract-load-transform (ELT), or change data capture (CDC). You do not need to memorize commands. Boomi Companion can build flows that bring in new data as well as flows that transform or organize data you already have loaded. The following examples show what you can ask and what you get back.
| Ask this | You get back |
|---|---|
| "List my BDI environments" | Every environment the token can see, with ids, so you can find the one you want. |
| "Give me a seven-day activity analysis of this environment" | Success rate, run counts and consumption, ranked top failures and top spenders, and a spend-spike check against the prior week. |
| "Find the last failed data flow and troubleshoot it" | The most recent failure, its root cause quoted from the run logs and configuration, a proposed fix, and how to roll it back. |
| "What connections do we have, and are any broken?" | The connection inventory with each connection's validity and last test error, so stale credentials surface immediately. |
| "Which tables can I replicate from this MySQL source?" | The source's tables with their incremental-cursor columns, so you can select a cursor before building the flow. |
| "Duplicate this data flow" | An exact copy of an existing flow, created disabled so nothing runs until you activate it. |
| "Build a flow replicating Postgres into Snowflake nightly" | A new data flow drafted from your description, reviewed with you before Boomi Companion creates anything. |
| "Where is our Boomi Data Units (BDU) consumption going, and where can we cut?" | BDU usage ranked by flow, with concrete savings opportunities, such as dead schedules, chronic failures, and full loads that should be incremental. |
A duplicated data flow is always created disabled, and it keeps the original's schedule. If you activate the copy without checking, it can start running on that same schedule and write to the same destination as the original. Confirm it is ready before you activate it.
Step 3: Review before anything changes
Reading is free-flowing. Boomi Companion answers questions about environments, data flows, connections, and consumption without asking for approval first. Any request that changes something is shown to you and waits for your go-ahead before Boomi Companion applies it. Some changes carry more weight than others:
- Some changes cannot be undone once made, such as clearing stored data or replacing existing settings.
- Some changes can be undone but still affect live data, such as editing, copying, or turning a flow on or off.
- Running a flow uses part of your account's data processing allowance.
Step 4: What stays in the BDI console?
Boomi Companion keeps a few actions out of scope by design: deleting assets, user and account administration, token generation, and installing kits from the marketplace. If a request needs one of these actions, Boomi Companion tells you and directs you to the BDI console to complete it.
