Gumloop Pricing: Pro, Credits and Enterprise Costs

Gumloop Pro starts at $37/month with 20,000 credits. Compare Enterprise, the 14-day trial, orchestration fees, usage examples and uncertain Free status.

Gumloop pricing article cover
G
Gumloop
Pricing last verified: 11/4/2026
View profile

Gumloop Pro starts at $37 per month and includes 20,000 credits each month, according to the vendor's public pricing page checked on November 4, 2026. The same page lists an 8% orchestration fee, unlimited seats and agents, and a 14-day trial. Enterprise uses custom pricing. There is no clear current Free-plan listing on the main price table, even though an older Gumloop announcement and the supplied Toollers record describe one. The central buying question is therefore not simply the subscription amount: it is how many credits your actual agents consume, how frequently they run, and which administrative controls the team needs.

Quick verdict

Pro is the published entry point at $37/month with 20,000 monthly credits and a 14-day trial. Estimate run-level consumption before committing. Enterprise is a sales conversation for identity, governance, custom capacity and possible VPC deployment.

Key takeaways

  • Pro lists 20,000 included credits each month
  • Starting subscription is $37/month; Enterprise is quote-based
  • A published 8% orchestration fee and variable model or tool work affect cost
  • Unlimited seats does not mean unlimited concurrent runs or compute
  • Current pricing page does not list the older permanent Free tier
G
Gumloop Tested

Create AI agents and automated workflows that connect business apps, models, and data.

Best for: Operations, GTM, support, and marketing teams prototyping AI agents with multiple SaaS integrations and reusable automations.

4.0
Research-based
Features and capabilities4.4
Usability and implementation3.8
Pricing and value transparency3.8
Integrations, security and trust4.0
Key features
  • Enterprise SSO/SCIM, policy controls and audit logs
  • Hosted MCP server on Pro
  • 35+ AI model choices
  • Prompt-led agent builder
  • Team collaboration and shared credentials
Pros
  • Pro includes unlimited seats, teams, and agents for $37/month before extra usage, favoring collaborative pilot projects.
  • Supports many model choices and bring-your-own model API keys rather than a single fixed LLM provider.
  • A large MCP connector directory supports cross-application agent tasks without building every connector manually.
Cons
  • Variable token, tool, compute, and orchestration consumption makes monthly costs harder to predict than seat-only subscriptions.
  • Enterprise security controls such as SAML/SCIM, organization-wide policies, and audit logs are not part of Pro.

What Gumloop charges for

Gumloop is an AI agent and workflow automation platform. It helps a business connect information sources, let agents reason over work, and take actions using linked services. An agent can support lead research, internal data analysis, content operations, email triage or another repeatable task with several steps. This scope is different from a basic connector that moves one data field when an event occurs. Complex reasoning, repeated model calls and external tool operations make usage a meaningful part of the bill. The Toollers Gumloop profile places the product in workflow automation software and records the same $37 Pro starting point.

The vendor's price architecture has two visible tiers: Pro and Enterprise. Pro exposes a fixed starting subscription with included monthly credit capacity. Enterprise changes the contract and organizational feature set, with custom credits and limits. The vendor does not show an ordinary per-seat charge for either tier in the current comparison, and both are described as allowing unlimited seats. That is useful for a team that wants participation from several functions, but it does not remove usage metering. If many colleagues invoke an agent simultaneously, the number of executions and their work can still drive costs, and the published concurrency limits remain relevant.

Gumloop also emphasizes selecting between AI models and connecting business applications. A task that requires simple text classification may cost less than a workflow that opens several pages, extracts information, sends it to a model, reasons again and writes records to multiple destinations. Some workflows have the same starting trigger yet very different consumption because the documents are longer, the model is more expensive, retries occur or the process calls another tool. A responsible budget should start with observed runs and an allowance for variation, even when the number of employees using the platform is easy to predict.

Gumloop pricing plans and included limits

PlanPublished base price in USDIncluded creditsKey access and limits
ProStarts at $37/month20,000/monthUnlimited seats, teams and agents; 8% orchestration fee; 14-day trial; one hosted MCP server
EnterpriseContact salesCustom allowanceOrganization-wide guardrails, custom capacity, SAML/SCIM, RBAC, audit logs; optional VPC

This chart reflects the main pricing comparison as of November 4, 2026. Gumloop does not provide a reliable annual-billing discount or a published Enterprise monthly dollar amount in that table, so neither appears as a fabricated annual equivalent. The $37 amount is explicitly a starting price. Specific subscription terms, credit top-ups and regional taxes should be confirmed in a live quotation or checkout screen. A volume or enterprise negotiation may change the rate and available controls, but no one should plan around an undocumented discount.

Pro: the starting plan for business agents

Pro lists 20,000 included credits each month. The current feature comparison includes unlimited agents, seats and teams, support for more than 35 models, bring-your-own model API keys, shared credentials, agent-scoped connector guardrails, one hosted MCP server and AI spend insights. Those are meaningful capabilities for a small team that wants several people to design or use automation instead of concentrating everything in one account. They do not establish that every connection is free to operate, that every external service is included, or that an AI agent can run without limits.

The same published table names five concurrent workflow runs in its legacy workflow area and 25 concurrent agent chats. These are throughput controls, not monthly quotas. An agency may have only twenty people, but if every person launches several long-running operations at once, concurrency can become the limiting factor before the credit allowance runs out. A large back-office batch may behave differently from an interactive assistant serving several employees. Ask whether queued work waits, fails or requires a change in plan, then test that behavior against a realistic peak in the trial.

The Pro support line includes Gumloop University and email support. An important procurement distinction is between access to learning materials and a guaranteed response time for production incidents. The public table does not give an ordinary Pro service-level agreement for incident handling. If a workflow updates financial records or triggers customer communication, the cost of a broken run may exceed the software subscription. Define a fallback route and an owner who can halt or repair actions; do not equate a low starting fee with included managed operations.

Enterprise: price and controls negotiated with sales

Enterprise has custom pricing and credit capacity, with more controls for organization-wide deployment. The feature table names role-based access control, SAML/SCIM support, an admin dashboard, audit logs, custom data retention rules, security reports, organization-wide policies, model access control and optional VPC infrastructure. Enterprises with internal IT review, regulated data, separate departments or central identity policies may find these requirements more decisive than the base cost. A team should ask sales to demonstrate the exact control needed rather than assume every advertised enterprise checkbox is enabled in every contract.

Gumloop describes unlimited agents and seats on Enterprise as well. However, custom limits apply to included credits, hosted MCP servers, concurrent legacy workflows and agent chats. Some deployment choices are optional. A quote should identify what is actually included, whether data export and retention controls require separate onboarding, which cloud environment hosts a private deployment, and who is responsible for support during an incident. A vendor statement that a VPC option exists is not evidence that a self-managed deployment can be provisioned without additional security and implementation work.

Dedicated Slack support is listed under Enterprise and an embedded Gumloop expert is shown as optional. Ask whether the contract covers after-hours incidents, hands-on initial development, security questionnaires or periodic usage reviews, and what each costs. For a procurement team comparing Pro with Enterprise, the real difference may be governance overhead avoided, support terms and the ability to control exposure to tools and models. Those benefits are difficult to express honestly as a simple price per seat, because the public offer does not sell them in that unit.

How Gumloop credits and the orchestration fee work

Credits are the unit used to measure and cover work. Gumloop's published explanation of model routing breaks an AI agent run into model-token usage, tool calls, compute and orchestration. The article describes a standard credit metering value and an 8% orchestration fee, and the public pricing comparison repeats that 8% for Pro. This makes Gumloop more variable than a flat-cost workflow manager. The exact amount for one run cannot be inferred from the agent's title, number of connected apps, or number of team members alone.

Consider a lead-enrichment agent. It could receive a list, search for publicly available company details, deduplicate a record, ask a model to summarize, then update a CRM. Each phase can do work, and the overall credit cost can change with the number of companies, source complexity, chosen model and retries. The difference matters even if the agent produces one tidy spreadsheet row in the end. A price estimate that multiplies only the number of final rows by an arbitrary fixed fee will miss the work the system performed to create them.

Bring-your-own API keys is listed on Pro and may provide model choice or billing flexibility. It also creates a need to distinguish charges made by the model vendor from those charged by Gumloop. The public comparison does not establish that using an outside key eliminates all Gumloop credits or the orchestration percentage. Ask for a line-item usage view under the exact credential arrangement you intend to adopt. Similarly, external paid data providers, sending services and connected applications can have independent subscriptions, request limits or usage-based fees that are not included in Gumloop's headline starting price.

The company describes spend insights, model controls and budgets. These matter because a successful automation can be run far more often than planned. A routine triggered on each new form submission may increase rapidly after a campaign launches. Set a spending threshold, review the cost of several ordinary runs, and define how a team should respond when an agent reaches its budget. For an agent that can update or send customer data, add an approval or testing checkpoint where the potential error cost is high. Cost and quality controls belong in the same operational design.

Three illustrated usage budgets

Scenario one: a cautious evaluation

Assume a pilot agent consumes 200 credits per completed run. At that hypothetical rate, a 20,000-credit monthly allowance would cover 100 such runs: 20,000 divided by 200 equals 100. If the team runs it three times on each of 22 workdays, it would execute 66 times and theoretically consume 13,200 credits. That leaves 6,800 credits for refinement, additional tests or another agent. The rate of 200 is an illustrative assumption, not a measured Gumloop benchmark. A real estimate requires an observed credit total for the exact models, documents and connectors.

The subscription would still begin at the advertised $37 per month before any applicable usage or external costs. We cannot compute a trustworthy total invoice from the hypothetical 13,200 credits alone because the accessible pricing table does not state the complete overage schedule, and the 8% orchestration charge must be understood in the actual usage statement. During the trial, record the billed credit and monetary line items instead of multiplying a marketing figure into an invented invoice.

Scenario two: a heavier document workflow

Suppose a document-processing agent requires 500 credits for each job in a sample estimate. The same monthly allowance covers 40 jobs: 20,000 divided by 500 equals 40. Eighty jobs would require 40,000 credits under that assumption, or twice the included allowance. The next step is not to announce a precise overage price. It is to inspect Gumloop's purchase screen or request a usage quote, because top-ups, pay-as-you-go settings, included credit accounting and orchestration can influence the final amount. If the model or document size changes, the assumed 500 credits per job also changes.

For a small operations team, the practical lesson is to pilot both typical and unusually difficult inputs. A short PDF with clearly labeled fields and a scanned document with inconsistent structure may lead to very different work. Test a sample that includes missing data and errors, and measure whether retries or validation loops create extra cost. A median-cost run is useful, but it is not a sufficient budget when bad inputs occur regularly. Record at least a typical, low and high consumption case before setting an alert.

Scenario three: a cross-team production rollout

Now consider sales, support and marketing all using shared agents. The published unlimited-seat claim means a simple per-person multiplier does not describe the Gumloop subscription. Instead, list each department's active agents, expected executions per day and credit consumption by task. If three departments each consume 15,000 credits monthly, the combined workload is 45,000 credits, above Pro's 20,000-credit starting allowance. That arithmetic identifies a potential capacity gap but still cannot price the excess without the applicable commercial rules. It also says nothing about concurrency peaks, which require a separate calculation.

This is where Enterprise may become relevant even when the organization does not need a large number of seats. Central administrators may want audit trails, role restrictions, identity integration, spending limits and separate connector permissions. A sensible request to sales would include expected monthly credits, concurrent agent chats, peak legacy workflow runs, desired private-network arrangement, and the need for named support contacts. The quote should show included and excluded charges as well as the process for raising limits. The buyer can then compare a predictable negotiated capacity against the risk of variable operational spending.

Is Gumloop free, and what does the trial include?

The current official Gumloop pricing page prominently says "Start 14-day free trial" for Pro, but it does not show a permanent Free plan alongside Pro and Enterprise. That conflicts with the Toollers product export, which still says "freemium" and marks a free plan available. An older official Gumloop article about simplifying its plans did describe a Free tier with monthly credits. Because a historical announcement is not a current checkout promise, this guide treats the 14-day trial as the verified present offer and leaves permanent Free-plan availability unresolved pending an actual signup check.

The trial is an opportunity to create a test agent, examine available connectors, check the visible usage ledger and learn which limits apply. Before entering a card, review the checkout information for whether payment details are required, what happens on day fifteen, whether credits or trial resources expire, and how to cancel. The publicly readable price grid does not establish all those terms. A team should use the trial to document plan fit and costs; it should not rely on an old blog post as proof that it can operate indefinitely at no charge.

Billing, credit rollover, cancellation and other questions

We did not find a current, definitive public annual-discount figure in the main pricing grid, so there is no annual total in our chart. The visible FAQ lists questions about unused-credit rollover and exceeding 20,000 credits in a month, but accessible page text does not provide an answer. This is a material budget variable. Ask whether credit allowances reset, whether unused capacity carries forward, how automated purchases work, and whether administrators can block further usage. Record Gumloop's answer in the plan agreement or support correspondence before authorizing large workloads.

Likewise, the public grid does not provide a universal refund or cancellation promise sufficient for us to assert a fixed refund window. Verify renewal timing, notice periods, treatment of prepaid capacity, what happens to agent definitions at cancellation and whether data exports remain available afterward. A company relying on agents for operations needs an exit plan that includes credentials, data, task histories and run configurations. That migration work can be more expensive than one or two months of subscription fees if it is deferred until a plan expires.

Taxes and payment processing depend on the buyer's location, so the advertised US-dollar starting price should not be mistaken for an after-tax invoice. The support model may also change the cost of ownership. Pro lists email support, whereas Enterprise lists dedicated Slack support and optional embedded expertise. Before signing a contract, ask which support channel covers outages, who takes responsibility when an integrated service's API changes, and what response target applies. Without those terms, the cheapest subscription might be poor value for a workflow that requires immediate recovery.

Who should choose Pro or Enterprise?

Who should use this

  • Small teams piloting AI agents with a measured credit budget
  • Operations teams that need agents spanning documents and connected applications
  • Businesses able to monitor run costs and handle exceptions
  • Larger organizations that need Enterprise identity, audit and access policies

Who should avoid this

  • Buyers who need an unconditional fixed price per workflow without credit metering
  • Teams that require confirmed permanent free operation before signup
  • Organizations that cannot permit autonomous system actions or control connector credentials
  • High-throughput deployments unwilling to validate concurrency and overage terms

Pro looks most attractive for a team that can start with a defined use case and assign one person to monitor quality, permissions and credit usage. A larger team could still use Pro if the needed access controls and throughput fit its published limits. Enterprise is appropriate when security teams demand identity integration, organizational guardrails, data retention controls or private deployment. The vendor's claim of unlimited seats is not a substitute for comparing the value of administrative controls or the risk associated with unattended automation.

Pros

  • Starting Pro price and included credit allowance are public
  • Unlimited seats, teams and agents simplify access planning
  • Multiple models and BYOK create configuration flexibility
  • Trial enables a bounded procurement evaluation

Cons

  • Total run cost varies with model and tool usage
  • Orchestration fee and possible outside API charges add complexity
  • Free-plan availability conflicts with older information
  • Concurrency and hosted-MCP limits still matter

Value compared with other automation tools

Pabbly Connect is a closer candidate for repetitive app-to-app triggers, actions and business integrations when advanced agent reasoning is not the central requirement. i10X is worth examining for a broader AI workflow builder. Apify and Browse AI can be better aligned with specialized web data extraction and monitoring, although neither automatically replaces Gumloop's collaborative agents and governance. These are existing Toollers profiles; the billing units and feature coverage differ enough that advertised entry prices alone cannot settle the choice.

To estimate value, calculate cost per accepted result, not simply cost per run. An agent that produces a completed, accurate customer record may save useful time; an agent that creates duplicate leads or sends incorrect messages can produce expensive cleanup work. Track successful outputs, human correction, rate of failed runs, cost per result and the burden of maintaining connectors. Include manual review and exception handling in the business case. A high-credit run that replaces substantive manual work can be worthwhile; a low-credit run producing unusable data is not cheap in practice.

Sources and pricing-check method

We checked Gumloop's public pricing comparison, its product and enterprise descriptions, and its published explanation of model-related usage on November 4, 2026 (UTC date). We also read its earlier pricing announcement to explain why a Free plan still appears in older records. We did not purchase a subscription, inspect a private billing dashboard or receive a contractual quote for this article. Pricing remains a source check, not a claim that a human editor verified a checkout invoice. The current Toollers export should be reconciled before its free-plan badge is treated as reliable.

Before spending money, request a written answer to four questions: what happens after the included credits, whether unused credits roll over, whether bring-your-own model keys change platform metering, and what cancellation does to agents and data. Then run representative low, normal and high-complexity tasks during the trial, retaining actual credit statements and screenshots. That evidence can convert an illustrative usage model into a usable departmental budget. No exact rate derived from another customer's workload should replace measurements under your own workflow.

Check the live Pro trial and ask Gumloop about credit overages

View Gumloop pricing

Frequently asked questions

How much does Gumloop cost?
Gumloop currently advertises Pro starting at $37 per month with 20,000 included monthly credits and an 8% orchestration fee. Enterprise has custom pricing. Actual costs depend on models, tool calls, compute and usage beyond the included credits.
Does Gumloop currently offer a permanent free plan?
The current main pricing page lists a 14-day Pro trial and Enterprise rather than a permanent Free plan. An older Gumloop announcement described a Free tier, but its present availability should be confirmed during signup.
How long is the Gumloop free trial?
The current public pricing page advertises a 14-day free trial of Pro. Check the signup and billing screens for trial terms and any payment-method requirement.
How many credits are included with Gumloop Pro?
The public plan comparison shows 20,000 credits per month in Pro. The amount consumed per run depends on AI models, tool operations and computing.
Are Gumloop agents and team seats limited on Pro?
Gumloop lists unlimited agents, seats and teams on Pro, but operational limits still include 25 concurrent agent chats, five concurrent legacy workflow runs, one hosted MCP server and credit-based usage.
How much does Gumloop Enterprise cost?
Enterprise is priced by sales quote. Its package can include custom credits, organization-wide policies, access controls, audit logs, identity integration and an optional VPC deployment.
Can Gumloop users bring their own AI API keys?
Yes. Gumloop lists bring-your-own API keys on Pro. The effect on individual provider charges and platform credit consumption should be checked before estimating a total bill.
Do unused Gumloop credits roll over?
The current public pricing page includes this question in its FAQ, but the accessible page text does not provide a definitive policy. Ask Gumloop support or inspect current checkout terms instead of assuming rollover.

Written by

Aditya Verma

Author

Toollers editorial contributor. Biography and credentials pending confirmation.

Reviewed by Shikha Goyal

We use optional analytics to understand site usage. No analytics loads until you consent.