Inngest pricing in 2026: what the free tier and $99 Pro cost at your volume

Short answer, checked on inngest.com/pricing on 8 October 2026: Inngest pricing in 2026 starts with a Free plan at $0 and 50,000 executions a month. Pro starts at $99/month, Business at $499/month, and Enterprise is custom. The catch is that every step counts as an execution, so the free quota covers fewer runs than the number suggests.

We build a pay-per-use scheduler that sometimes competes with Inngest, so read the section on scheduled HTTP calls near the end with that in mind. Everything before it is Inngest's own numbers and our arithmetic.

Inngest pricing at a glance

Full details are on Inngest's pricing page.

If you read a pricing explainer from July 2026, two things are different now. The free plan is called Free, and there's a Business tier at $499. There's still nothing between $0 and $99.

What counts as an execution

Inngest's formula is "Executions = runs × (steps + 1)". Their own example: a function with 5 step.run() calls uses 6 executions per run. Their FAQ also says a multi-step AI agent can often use 10+ executions per run.

It's easy to read "50k executions" as 50k runs. That's only true for a function with no steps. Here's what the free quota covers as you add steps:

So one step halves your free runs, and an agent that wraps 10 tool calls in steps gets about 4,545 runs a month before the quota is gone.

What happens when the free tier runs out

The Inngest free tier in 2026 has no overage. In Inngest's words, "execution pauses once you hit the free quota, and you'll need to upgrade to resume." The next plan up is Pro at $99/month.

Executions aren't the only limit on Free, though. The plan also caps:

• 5 concurrent steps (this is the usual complaint)
• 24 hours of trace history
• 30 days max run length, including sleeps and waits
• 3 connected workers (serverless workers are unlimited)
• 5 seats

Any of these can push you to Pro while you're using a small part of the 50k executions. That's the case where $99 feels expensive: you pay for 1M executions because you needed a sixth concurrent step or more than 24 hours of traces.

What Pro really costs at your volume

Take a scheduled job with one step. That's 2 executions per run, on Pro's $99 base with 1M executions included:

At 30,000 runs this job is already past Free (60k executions is over the 50k cap), and Pro is still cheaper per run than a cent.

On top of the base price:

• Executions past 1M bill on a tiered overage, "from $0.000050 down to $0.000015 per execution".
• Concurrency above the 100 included steps is $25 per 25.
• Seats past 15 are $10 each.
• Events past 5M are $0.50 per 1M.
• Advanced observability (Datadog and similar) is listed as a $300 add-on, and HIPAA is an add-on too.

Each of these bills on its own, and Inngest's FAQ says this is why bills can vary month to month. Step-heavy runs are where it shows up first. With 10 steps per run, each run is 11 executions, so the 1M included covers about 90,909 runs instead of 500,000.

We went through the general version of this math (monthly floor against per-call price) in pay-per-use vs subscription pricing.

When Inngest is the cheap option

Most of the time, honestly. On price alone, Inngest beats $0.01 a run almost everywhere. Inside Free it costs nothing, and past the free cap Pro is cheaper per run than our $0.01, with the gap growing the more you run. Inngest is also open source, so you can self-host it and pay only for your own servers.

If you actually use durable steps, events and fan-out, $99 for that is a fair price. We compared the two models side by side in Omniyond vs Inngest. And if you've decided to leave anyway, how to migrate off Inngest covers the cutover step by step.

If you only needed a scheduled HTTP call

If you're paying $99 only for the concurrency cap, or for one scheduled job behind one feature, you might not need durable functions at all. Omniyond is a pay-per-use API toolbox for AI builders: hosted HTTP scheduling and utilities on one key and one balance, from $0.01 a call. A scheduled run is $0.01, packs are $9, $29 or $99 one-time and never expire, and there's no SDK: you call a URL, or add it to an MCP client with npx omniyond-mcp. The limits: recurrence is interval-only (5 minute minimum, 90 day maximum, no cron expressions or time zones), and there's no multi-step orchestration. If an agent is the one that needs to act later, start with cron for AI agents, and for other tools see our Inngest alternatives grouped by job.

FAQ

Is Inngest free?

Yes. The Free plan is $0, needs no credit card, and includes 50,000 executions a month, 5 concurrent steps and 24 hours of trace history.

What counts as an execution in Inngest?

Each run plus each step in it: executions = runs × (steps + 1). A function with 5 step.run() calls uses 6 executions per run.

What happens when I hit the Inngest free tier limit?

Execution pauses until you upgrade. Free has no overage billing.

How much is Inngest Pro, and what are the overage rates?

Pro starts at $99/month with 1M executions included. Overage is tiered, from $0.000050 down to $0.000015 per execution.

Is there a plan between Free and $99?

No. After Free comes Pro at $99/month, then Business at $499/month.

How many runs fit in the free tier?

50,000 if your function has no steps, 25,000 with one step, 12,500 with three, and about 4,545 with ten.

Can I self-host Inngest to avoid the bill?

Yes, Inngest is open source and can be self-hosted. You swap the subscription for running the servers yourself.

The short version

Free is generous if your functions have few steps and you stay under the other limits. Past that, Pro at $99 is cheap per run at any real volume, and the surprises come from step-heavy runs and add-ons. Prices change, so check inngest.com/pricing before you commit.

If all you need is one scheduled HTTP call, a $9 Omniyond pack is enough to try it.