Resources · Make integration

Braintrust Make integrationAutomate Braintrust with Make.

Can your evals run without anyone opening the Braintrust dashboard? The Braintrust Make integration gives you 4 modules: three actions and one search. This guide is for AI teams that want to connect the account, then build a first scenario that creates an experiment and launches an evaluation.

Try the Braintrust integration on MakeSee the 4 modules

Verified Trustpilot reviews · AI, automation & growth agency

What you can automate

What can the Braintrust Make integration do for you?

The Braintrust Make integration links your Braintrust account to Make, a visual tool where you build a scenario (the automation itself, a chain of modules) without writing code. Each module is one brick: here, creating an experiment, launching an evaluation, listing functions or calling the API directly.

An experiment opened for every new model version. When another tool you already use signals a change, Create a Experiment opens a fresh experiment in Braintrust, so results stay grouped under a clear name instead of piling up by hand.

An evaluation launched right after. Chain Update an Evaluation Module behind it: the module launches an evaluation as soon as the experiment exists, so no new version goes untested.

An inventory of what your project exposes. List Functions returns every function as a separate bundle (one item passed from module to module), which makes it easy to review them or copy them elsewhere.

What Make does not do here: Braintrust has no trigger module, so nothing starts on its own from inside Braintrust. The scenario begins on a schedule or on another app. For anything the three dedicated modules skip, Make an API Call reaches the rest of the API. The same app often exists as an n8n node too; the n8n vs Make page lays out the differences, and the Make pricing page explains what each plan allows. If you would rather learn the editor properly first, see the Make training.

Connect

How do you connect Braintrust to Make?

  1. 01

    Copy your API key in Braintrust

    In Braintrust, open the organization settings page, go to the API keys section and copy the key shown on screen. You need an active Braintrust account for this; the subscription itself is handled on the Braintrust side.

  2. 02

    Add a Braintrust module in Make

    Log in to Make, open your scenario, add any Braintrust module and click Create a connection. The connection is your Braintrust account linked to Make once, then reused by every Braintrust module you add later.

  3. 03

    Paste the key and save

    Give the connection a label in Connection name if you like, it is optional. Paste the API key you copied, confirm, and Make now lists what it can read from your account inside the module fields.

First scenario

Your first scenario with Braintrust

GoalWhen another app you use sends new data, Make creates an experiment in Braintrust and launches an evaluation.

  1. 01

    Create the scenario

    From the Scenarios page, create a new scenario and click the +. Pick the app that should start the flow, since Braintrust has no trigger of its own.

  2. 02

    Add Create a Experiment

    Click the + to the right of that first module, search for Braintrust and choose Create a Experiment. Select your connection or click Create a connection with your API key.

  3. 03

    Chain the evaluation

    Add Update an Evaluation Module after it and fill its fields. Map the data coming from the previous modules instead of typing values by hand, so each run uses fresh information.

  4. 04

    Test with Run once

    Click Run once. Make executes the scenario a single time and shows the bundles each module received, which is the fastest way to check that the experiment and the evaluation really happened.

  5. 05

    Schedule and switch it on

    Open the clock on the first module to set the schedule, every 15 minutes by default, then switch the scenario on. From then on, every run counts operations on your plan.

Modules

What the Braintrust modules do

Braintrust gives you 4 modules. For each one: what it does for you, when to reach for it, and what to watch out for.

Braintrust1

Create a Experiment

Action

In Make“Creates an experiment.”

You get a new experiment in Braintrust without clicking through the interface, ready to collect the results of the next test run.

When to use it
as the first Braintrust step after another app signals that a prompt or model version changed, so every version gets its own experiment.
Watch out
each run creates one more experiment, so test with Run once before scheduling the scenario.
Braintrust2

Make an API Call

Action

In Make“Performs an arbitrary authorized API call.”

This one opens the rest of the Braintrust API to your scenario, reusing the connection you already set up, with no second key to paste.

When to use it
when the endpoint you need has no dedicated module among the three others, rather than leaving Make for a separate script.
Watch out
you describe the request yourself, so keep the Braintrust API reference open while you fill it in.
Braintrust3

Update an Evaluation Module

Action

In Make“Launches an evaluation.”

Despite the word update in its name, the directory describes this module as the one that launches an evaluation, so a new round of evals starts without anyone opening Braintrust.

When to use it
right after Create a Experiment, or whenever an upstream event should kick off a new round of evals.
Watch out
the documentation is AI-generated and may contain mistakes, so confirm the behavior with one Run once.
Braintrust4

List Functions

Search

In Make“Lists all functions.”

Here the output is a full list of your Braintrust functions, delivered as one bundle per function so the next modules can handle them one at a time.

When to use it
to build an inventory in a spreadsheet or to check which functions exist before a scenario acts on them.
Watch out
a search can return zero, one or many bundles, and each bundle processed downstream uses operations.
Need help

Need help automating Braintrust with Make?

A person reads every message.

FAQ

Braintrust and Make: common questions

01Is the Braintrust Make integration free?
Yes, on the Make side. Braintrust is a standard app, so its modules work from the Free plan onward. That plan allows 2 active scenarios, a minimum of 15 minutes between two scheduled runs, 5 minutes of execution per run and 512 MB of data transfer. Paid plans (Core, Pro, Teams and Enterprise) drop the interval to 1 minute and remove the cap on active scenarios. Braintrust itself still needs an active account, and its subscription is set on the Braintrust pricing page, not in Make.
02What do you need to connect Braintrust to Make?
Two things: an active Braintrust account and an API key. The key sits in the organization settings page, under the API keys section, where you copy it. In Make, you add a Braintrust module, click Create a connection, optionally name the connection and paste the key. The connection is then shared by all four Braintrust modules, so this setup happens once per account rather than once per module or per scenario.
03Can Braintrust start a Make scenario on its own?
No. Braintrust has no trigger module in Make, neither scheduled nor instant, so nothing inside Braintrust can launch the scenario. Two ways around it. The first is the schedule on the first module: a new scenario runs every 15 minutes by default, with 15 minutes as the minimum on the Free plan and 1 minute on paid plans. The second is to open the scenario with a trigger from another app, then add the Braintrust modules behind it.
04What if the Braintrust module you need is missing in Make?
Use Make an API Call. It calls any endpoint of the Braintrust API with the same connection, which covers what Create a Experiment, Update an Evaluation Module and List Functions leave out. Keep in mind that the Make documentation for this app is AI-generated from the Braintrust docs and may contain mistakes. When a detail matters, check it against the Braintrust documentation or ask Braintrust directly, then validate the result with Run once before switching the scenario on.
05Should you use Make or n8n for Braintrust?
It depends on where the rest of your automations already run. Braintrust often has an n8n node as well, so the choice is less about the app than about the tool your team works in. On the Make side, you get four modules, including a generic API call for anything the others miss, and a visual editor where each step shows the bundles it received. Compare the two on your real use case rather than on a general ranking, since neither is better for every team.
Same platform

Other docs you might need on this stack.

Browse all integrations