Do you need to know how to code to build an agent on Gumloop?

Not for the essentials: flows are built on a visual canvas, node by node, connecting the software the team already uses. Advanced cases (custom scripts, in-house connectors) remain accessible to those who can code, without being required to get started.

Does my data train an AI model?

Not for paid ("premium") accounts: the privacy policy explicitly excludes submitted data, built flows and exchanges with agents from any training, including at OpenAI and Anthropic for calls made through Gumloop. The exact fate of data during the 14-day free trial isn't specified separately.

Where is my data processed?

In the United States by default, according to the privacy policy. Dedicated infrastructure in another region is only available on the Enterprise plan, on quote, not as a setting available on standard tiers.

How much does Gumloop cost for a team of ten?

The Pro tier is billed by credit volume consumed, not by headcount: $37 a month for about 20,000 credits, with unlimited seats. Ten people sharing the same flows therefore don't pay ten times more; heavy credit use does.

What's the difference with n8n or Zapier?

n8n and Zapier are first and foremost automation tools between software, with an AI step added to the flow. Gumloop starts from the opposite: the AI agent is the central piece, and connectors give it access to business tools. The choice depends on which matters more in your case: linking software together, or having an agent act on it.

Gumloop: our review

Gumloop builds AI agents on a visual canvas, connected to Slack, Gmail, Salesforce and about a hundred other business tools. No permanent free tier, only a fourteen-day trial before paying.

Data in the United StatesPaid·from $37/month

Go to Gumloop's website

What we take away from it

What counts in its favor

Written commitment not to use paid accounts' data, flows or exchanges to train a model, citing a direct agreement with OpenAI and Anthropic on this point.
A wide, concrete connector catalog (Slack, Gmail, Salesforce, HubSpot, Snowflake…), built to plug an agent into tools already in place rather than impose new ones.
Billed by credits consumed, not by seat: opening the tool to a whole team doesn't automatically multiply the bill.

What counts against it

No permanent free tier: only a 14-day trial precedes the first paid tier at $37/month.
Data processed in the United States by default; dedicated infrastructure (VPC) only exists on the Enterprise plan, on quote.
The no-training commitment is worded for 'premium' accounts: the documentation doesn't explicitly state what happens to data during the 14-day trial.
No security certification (SOC 2, ISO 27001) found publicly on the site as of our last check.

A good choice for a team that has already identified the software to connect and wants an operational agent fast, with no contract to negotiate. Skip it if the budget can't absorb a subscription from day one, or if data must stay outside the United States without an Enterprise negotiation.

The facts

First paid tier: $37/month — Pro, billed monthly (-20% annually)
Data hosting: Data in the United States · source
Categories: AI agents,Automation
Data verified on: 30 July 2026

What happens to your data

Is your data used to train their models?: No
Subprocessors: The privacy policy states: "For premium users, we do not use any of your uploaded data, flows, or chats with agents that you create to train any AI whatsoever", and adds a direct commitment with OpenAI and Anthropic on data sent through Gumloop's API. "Premium" refers to paid accounts (Pro and Enterprise): since Gumloop has no permanent free tier, this covers most real-world use, but the documentation doesn't explicitly say whether the 14-day trial is covered by the same clause. Processing happens in the United States by default ("we transfer the data, including Personal Data, to United States and process it there"); dedicated infrastructure (VPC) is only offered on Enterprise.
Retention: Kept "only for as long as is necessary for the purposes set out in this Privacy Policy"; no figure published.

Gumloop in detail

Gumloop builds AI agents connected to a company's software: a visual canvas where each node does one thing — read a Slack message, query a spreadsheet, write to the CRM — and where an agent decides, within that frame, how to handle each case.

What it actually does

A Gumloop flow starts from a trigger — a received message, a row added to a spreadsheet, a webhook — and chains together nodes that query third-party software or a language model. The difference from a classic automation tool lies in the agent's role: rather than following a fixed rule, it can decide which branch to take, write an original reply or flag a case that falls outside the frame.

The most common use case: an agent that reads incoming tickets, sorts them by urgency, enriches the customer's record in the CRM and drafts a reply, left for a human to approve before sending.

Why it interests us

Gumloop publishes a written commitment fairly rare in this category: paid accounts' data isn't used to train a model, and the vendor says it obtained the same commitment directly from OpenAI and Anthropic for calls made through its API. That's the kind of guarantee we look for before connecting an agent to a business mailbox or CRM, in our work on automation and AI agents. The reservation lies in location: processing stays American by default, and another region requires signing an Enterprise contract.

Where it falls short

First point: no permanent free tier. The trial stops at 14 days, after which you have to pay to continue — a different choice from most automation tools, which often leave a limited free tier available long-term.

Second point: the no-training commitment targets "premium" accounts, a term the privacy policy doesn't explicitly apply to the trial period. When in doubt, it's better not to run sensitive data through it before confirming the account's status.

Third point: no public security certification (SOC 2, ISO 27001) was found on the site at last check, even though the Enterprise page mentions "security reports" provided on request — so probably existing, but not published in open access.

Our recommendation

Gumloop suits a team that has already identified the software to connect and wants a working agent quickly, with no contract to negotiate to get started. For an organization that wants to test for a long time before paying, or that requires hosting outside the United States from the first tier, it's better to look elsewhere or negotiate the Enterprise plan directly.

Who it's for, and who it isn't

Suited for

An SME or internal team that wants to automate a recurring task (ticket sorting, CRM enrichment, meeting prep) without writing code
An organization already using Slack, Gmail or Salesforce, looking to plug an agent into them rather than switch tools
A team ready to pay from the start: the model relies on a short trial, not extended free use

Not a fit

An organization that wants to test for a long time before paying: the trial stops at 14 days
An organization that requires hosting outside the United States without going through an Enterprise contract
A need for a public, verifiable security certification before any rollout

Frequently asked questions

Do you need to know how to code to build an agent on Gumloop?: Not for the essentials: flows are built on a visual canvas, node by node, connecting the software the team already uses. Advanced cases (custom scripts, in-house connectors) remain accessible to those who can code, without being required to get started.
Does my data train an AI model?: Not for paid ("premium") accounts: the privacy policy explicitly excludes submitted data, built flows and exchanges with agents from any training, including at OpenAI and Anthropic for calls made through Gumloop. The exact fate of data during the 14-day free trial isn't specified separately.
Where is my data processed?: In the United States by default, according to the privacy policy. Dedicated infrastructure in another region is only available on the Enterprise plan, on quote, not as a setting available on standard tiers.
How much does Gumloop cost for a team of ten?: The Pro tier is billed by credit volume consumed, not by headcount: $37 a month for about 20,000 credits, with unlimited seats. Ten people sharing the same flows therefore don't pay ten times more; heavy credit use does.
What's the difference with n8n or Zapier?: n8n and Zapier are first and foremost automation tools between software, with an AI step added to the flow. Gumloop starts from the opposite: the AI agent is the central piece, and connectors give it access to business tools. The choice depends on which matters more in your case: linking software together, or having an agent act on it.