We made Hatchet cheaper to run. Now it’s cheaper to use.
In short
- Self-serve Hatchet just got a lot cheaper: pure pay-as-you-go with 10× more free usage and no plan fees.
- Add a card and your first 1 million task runs and 1 million events each month are free, up from 100k. After that, it’s $30 per million task runs and $5 per million events.
- We’re moving Developer plans to pay-as-you-go and retiring Team, Starter, and Growth plans. Most of you will pay less. A lot of you will pay nothing.
What’s changing?
We’re going from four plans to two:
Pay-as-you-go: Add a card and you get up to 1M runs and 1M events free every month, then usage-based, plus 10 tenants, 50 users, 100 workers, 10,000 worker slots, 1,000 cron jobs, and 7-day retention.
Custom: Volume discounts for teams running 100M+ runs a month, as well as dedicated capacity, multi-region, SLAs, SSO, HIPAA, audit logs, self-hosted support, and a shared Slack channel with our engineers.
The full breakdown is on the pricing page.
How does this compare to other workflow orchestrators?
Thank you for asking! This is the part we’re most excited about, because we think workflow orchestration should be a cheap and reliable part of your stack.
Hatchet and Temporal count work in a similar way: a workflow and each task or activity therein are billed as a unit. The differences are cost and retries. Hatchet’s first 1M runs each month are free, then it’s $30 per million, and retries aren’t billed. Temporal charges $50 per million actions from the first one, plus a 10% support fee, and bills retries, timers, signals, heartbeats, and other operations as separate actions. DBOS’ Pro and Teams plans charge per checkpoint (every workflow and every step), and on those plans you bring and pay for your own Postgres.
Here’s a comparison of a sample workflow’s monthly costs:
We’re using an example workflow that represents an agentic loop running 10 child tasks. Each step is set up as its own task, activity, or step, following each product’s docs. This works out to 11 runs per workflow on Hatchet, 11 actions on Temporal, and 11 checkpoints on DBOS.
†At 10M workflows (110M task runs a month) this is the pure pay-as-you-go list price. Hatchet’s Custom plan offers discounts on this volume, so your bill would likely be lower.
We’re also not including storage costs. Hatchet charges for payload storage at $0.03 per GB-hour after 10 GB of active storage per day free on pay-as-you-go. Temporal’s pay-as-you-go plan includes no free storage and charges $0.042 per GB-hour on each workflow’s full event history from the first GB, plus retained-storage fees after a workflow finishes.
*At 5M and 10M workflows, Temporal’s “Business plan” (volume pricing, $500/month minimum, 2.5M actions included) is cheaper, so we used that. As mentioned above, storage isn’t included in this estimate. Also, Temporal retries, timers, signals, and heartbeats would add more actions, but this example doesn’t use them.
**We used DBOS Pro ($99/month, 1M checkpoints included) up to 250k workflows, then DBOS Teams ($499/month, 10M checkpoints included) where it’s cheaper. DBOS also needs a Postgres database you host yourself.
These estimates were created in good faith based on each product’s cheapest published plan at list prices as of September 2026. Negotiated and commitment discounts aren’t included for anyone, including Hatchet. Temporal’s published volume pricing is included where it’s cheaper. If we got something wrong here, let us know and we’ll fix it. (Sources: Temporal pricing, Temporal billable actions, DBOS pricing.)
Why are we doing this?
Because Hatchet got a lot cheaper for us to run and we’d rather hand those savings to you than keep them.
A year ago, Hatchet Cloud ran on a handful of very large database shards in one region. Today, it runs on 20+ smaller shards across 4 regions, with more added every month. Smaller shards mean we can size resources to the actual load instead of paying for giant machines that sit idle.
We’ve also spent two years obsessing over performance. Faster code means less load on the database, fewer spikes, and fewer machines. That’s what makes Hatchet more reliable, and it’s also what makes it cheaper.
… But won’t you make less money?
Yes, we will make less money in the short term. Some of our customers will pay a lot less and some will pay nothing at all, and we’re okay with that. We think more and more of the software people build is going to run asynchronously: agents, background jobs, pipelines, and more. That work needs a reliable and efficient place to run and we want Hatchet to be the obvious default. The best way for us to get there is to make it as easy (and cheap!) as possible to start.
We’re confident this is also good for our business. Teams that start on Hatchet tend to stay and grow with it. So we’d rather remove every reason not to try it.
How will this affect me?
While plan fees are going away, the price per run after your free million is increasing, from $10 to $30 per million. If you run tens of millions of task runs a month, your bill may go up. Our Custom plans are designed to offer volume discounts if you’re running a high-volume workload. If you have any questions or concerns, talk to us and we’ll work it out.
Do I need to do anything?
- Not on a plan yet? The updated pricing is live today for all new signups.
- On a Free plan? You’ll move to pay-as-you-go with no-card-on-file limits. Most limits stay the same or go up. Add a card to get 1M free task runs a month and higher limits.
- On a Developer plan? You’ll move to pay-as-you-go automatically.
- On a Team, Starter, or Growth monthly plan? You can stay on your current plan through the end of this year or switch to pay-as-you-go at any time. For most accounts on these plans, pay-as-you-go will be cheaper. We’ll reach out to help you compare.
- On a pre-paid annual Team, Starter, or Growth plan? Stay on your current plan until it expires or switch to pay-as-you-go at any time (we’ll refund the prorated difference or apply the value toward PAYG).
- On a Scale plan or a contract? Nothing changes.
What should I do now?
Keep building. If you’ve been holding a workload back because of cost, a million free runs a month is a lot of room. Get started.
Questions? Talk to us or find us on Discord.