Comparison
Gumloop pricing explained: plans, credits, and real costs
Direct answer
Here is Gumloop pricing as of October 2026. Gumloop has two plans, per Gumloop's pricing page: Pro, which starts at $37 a month with a 14-day free trial and 20,000 credits a month, and Enterprise at custom pricing. Credits are worth $0.005 each and pay for the list cost of AI model tokens, compute, and paid tool calls, plus an 8% orchestration fee. That fee rises to 16% if you bring your own model API key. Unused credits don't roll over, and once you run out, agents stop until you turn on pay-as-you-go at $0.005 a credit.
Gumloop plans at a glance
Figures are from Gumloop's pricing page as of October 2026.
| Pro | Enterprise | |
|---|---|---|
| Price | From $37/month, 14-day free trial | Custom |
| Included credits | 20,000 a month (7,400 plus 12,600 bonus) | Custom |
| Orchestration fee | 8% (16% with your own model key) | 8%, discounts available |
| Agents | Unlimited | Unlimited |
| Models | 35+ and custom proxy support, bring your own API keys | Same |
| Seats and teams | Unlimited | Unlimited |
| Shared credentials | Yes | Yes |
| Concurrent workflow runs | 5 | Custom |
| Concurrent agent chats | 25 | Custom |
| MCP server hosting | 1 | Custom |
| Support | Gumloop University, email | Adds dedicated Slack support, optional embedded Gumloop expert |
| Admin and security | Agent-scoped connector policies | Org-wide policies, role-based access, SCIM/SAML, audit logs, admin dashboard, custom data retention, optional virtual private cloud |
The page now leads with agents. Gumloop's original node-based workflows are labeled "Workflows (Legacy)," though Pro still allows 5 concurrent workflow runs.
How Gumloop credits work
Gumloop uses credits as one currency across model providers and the 30+ paid services it connects to. The pricing page FAQ explains the math. For every agent run, Gumloop bills the list cost of:
- Tokens from model providers, at each provider's list price
- Compute from compute providers, at their list prices
- Tool calls to paid APIs such as Apollo, Exa, or Parallel, at each provider's list price, with a base cost of 1 credit per call
On top of that total, Gumloop adds an 8% orchestration fee.
A few details matter for your bill:
- The $37 buys more than $37 of usage. At $0.005 a credit, 7,400 credits equal $37. The 12,600 bonus credits bring the monthly total to 20,000 credits, or $100 of usage at list price.
- Credits reset monthly. The 20,000-credit grant renews each billing period, and unused credits don't carry over.
- Running out stops your agents. Pay-as-you-go is off by default. When you pass 20,000 credits mid-month, agents stop working and Gumloop prompts you to turn it on. After that, extra usage bills at $0.005 a credit.
Bringing your own API key
Pro and Enterprise let you plug in your own keys, which changes the math:
- Your own model key: the credits that would have paid for tokens drop to zero, since you pay the model provider directly. The orchestration fee doubles from 8% to 16%, and it's calculated on what the task would have cost at list price without your key.
- Your own key for MCP tool calls: credit costs fall to a minimum of 1 credit per tool call.
So bringing your own model key makes sense when you get better rates from the provider than its list price, or when you already have committed spend there. If you pay list price anyway, you're paying the provider for tokens plus a 16% fee to Gumloop, instead of 8%.
A worked example for a small business
These numbers are an illustration, not a benchmark. Model prices vary widely, so the per-run cost here is an assumption.
Say a 12-person marketing agency runs a lead research agent. Each run looks up a company, pulls contact data from a paid API, and writes a short brief. Assume the list cost of one run, counting tokens, compute, and tool calls, is $0.20.
| Step | Calculation | Result |
|---|---|---|
| Cost per run with the 8% fee | $0.20 x 1.08 | $0.216, or 43.2 credits |
| Runs covered by the Pro plan | 20,000 / 43.2 | About 460 a month |
| 800 runs a month | 800 x 43.2 = 34,560 credits | 14,560 credits over |
| Overage on pay-as-you-go | 14,560 x $0.005 | $72.80 |
| Monthly total at 800 runs | $37 + $72.80 | $109.80 |
Now the same 800 runs with your own model key, assuming tokens make up $0.12 of the $0.20:
- Gumloop credits per run cover the remaining $0.08 of tool calls and compute, plus a 16% fee on the full $0.20 ($0.032). That's $0.112, or 22.4 credits.
- 800 x 22.4 = 17,920 credits, which fits inside the Pro plan.
- You pay $37 to Gumloop plus about $96 (800 x $0.12) to the model provider at list price, so roughly $133 a month in total.
In this example, bringing your own key costs more unless your provider rate is well below list. The general lesson holds for any Gumloop bill: the model you choose drives most of the cost, and a heavier model can multiply the per-run price quickly.
Hidden costs beyond the plan
- Building and tuning. You design each agent, connect tools, write instructions, and test outputs. Better prompts and cheaper models lower the bill, but that work takes time.
- Model choice. Since credits pass through list token prices, switching to a larger model can raise costs several times over with no change to the plan.
- Stopped agents. With pay-as-you-go off, agents stop when credits run out. If those agents handle anything time-sensitive, turn it on or watch usage closely.
- Judgment and review. Agents can draft and research well, but someone still checks what goes out to customers.
When Gumloop is the right buy
Gumloop is a good fit when:
- You have someone who likes building AI agents and wants a wide choice of models.
- Your team wants one shared pool of credits and unlimited seats at a low entry price.
- Your work is research, data enrichment, or content-heavy tasks where AI agents do well.
- You want token costs passed through at list price instead of bundled into opaque units.
It's a weaker fit if nobody has time to build and maintain agents, or if usage is hard to predict and you can't watch the credit meter.
Alternatives to consider
- Automation tools like Zapier, Make, and n8n are better for rule-based steps between apps, where an AI model isn't needed on every run.
- Other agent builders. See our Lindy alternatives roundup.
- A done-for-you team. TeamShift is a hosted team of nine AI workers that knows how your business runs. You don't build agents. You hand off work in a sentence, and the right worker, such as Research or Sales, does it in the apps you already use. Workers draft outreach, replies, and posts, and anything that sends to a customer, moves money, publishes, deletes, or commits your calendar goes out only after you approve it. Pricing is usage-based: a free workspace, work billed at 3x its compute cost, a quoted maximum before each job, and daily or monthly spend caps (pricing).
Pick Gumloop if you want to build your own agents with a broad model menu at a low starting price. A done-for-you team makes more sense when you want the work done without designing and maintaining agents yourself.
FAQ
How much does Gumloop cost?
As of October 2026, Gumloop Pro starts at $37 a month with 20,000 credits and a 14-day free trial. Enterprise pricing is custom.
Is there a free Gumloop plan?
The pricing page lists Pro and Enterprise, with a 14-day free trial of Pro. There's no ongoing free tier shown on the page as of October 2026.
What is a Gumloop credit?
A credit is worth $0.005 and pays for model tokens, compute, and paid tool calls at list price. Gumloop adds an 8% orchestration fee on top, or 16% if you use your own model key.
What happens when I run out of Gumloop credits?
Your agents stop working until you turn on pay-as-you-go, which is off by default. Once it's on, extra usage costs $0.005 a credit.
Do unused Gumloop credits roll over?
No. The 20,000-credit grant resets each billing period, and you don't bank unused credits.