Pricing

machine.dev bills per minute in US dollars. You only pay for runtime. Provisioning and teardown are free. There are no credits, no commitments, and the only thing on your bill is the seconds your runner is up plus the storage it uses.

Prices below come from the canonical pricing API and reflect the best rate across all regions. Spot rates fluctuate; the live pricing page is always authoritative.

How billing works

• Per-minute, USD. Your bill is runtime in minutes × per-minute rate for compute, plus prorated EBS storage charges.
• Pay-as-you-go is the default. No subscription required.
• Subscription plans bundle a bonus (see below).
• Spot rates change with the AWS market. You’re charged the rate at job start.
• No charge for provisioning. Cold start takes ~1 minute and is free.
• No charge for failed provisioning. If we can’t get you a runner, you pay nothing.
• Spot interruptions are billed up to the moment of interruption, then stop.
• The dashboard shows dollars by default. You can toggle to credit-style accounting in Settings if you prefer.

Free tier

New accounts receive $10 of free compute on signup. At spot rates that’s roughly:

• ~63 hours of T4G GPU time
• ~37 hours of L4 GPU time
• ~140 hours of CPU time (4-vCPU ARM)

Non-paying teams can run one job at a time, with a 30-minute limit per job. Buying compute or subscribing removes the 30-minute limit. Promotional grants alone do not count as a purchase. See Billing & limits for paid concurrency limits.

Sign up to claim it.

CPU runners

X64 (Intel/AMD)

ARM64 (Graviton)

GPU runners

Each NVIDIA GPU is offered in several vCPU/RAM configurations; RTX 6000 also comes with 2, 4 or 8 GPUs. Pick a higher-vCPU config if your workflow does heavy CPU-side preprocessing.

AWS AI accelerators

EBS storage pricing

Storage is billed separately from compute and prorated to your job’s runtime. All runners use gp3 EBS volumes.

Costs are prorated per minute of runtime. The default 6,000 IOPS and 250 MB/s throughput are included at no additional charge. Only IOPS and throughput above those defaults incur extra charges.

Default storage configuration:

• Volume size: 100 GB
• IOPS: 6,000
• Throughput: 250 MB/s

Example: default storage on a 30-minute job

• Storage: 100 GB × $0.08 / 43,200 min × 30 min = $0.006
• IOPS surcharge: none (default)
• Throughput surcharge: none (default)
• Total EBS cost: ~$0.006

Example: 500 GB / 10,000 IOPS / 750 MB/s on a 60-minute job

• Storage: 500 × $0.08 / 43,200 × 60 = $0.056
• IOPS surcharge: (10,000 − 6,000) × $0.005 / 43,200 × 60 = $0.028
• Throughput surcharge: (750 − 250) × $0.04 / 43,200 × 60 = $0.028
• Total EBS cost: ~$0.112

Use the disk_size, disk_iops, disk_throughput labels to customise per job.

Subscription plans

Plans are an optional bonus on top of pay-as-you-go. Each tier gives you more dollar value than you put in. All compute is still billed per-minute in USD. Plans provide a bonus balance and a higher concurrent-runner limit. Buying pay-as-you-go compute unlocks five concurrent runners without a subscription.

Pick a plan in the machine.dev dashboard or contact sales for enterprise pricing.

Jobs above your team’s cap wait in arrival order for up to six hours. Cancelling a waiting job removes it from the queue. See Billing & limits for runtime limits, auto top-up and invoices.

FAQ

What happened to credits?

machine.dev now bills in US dollars instead of credits. If you had a credit balance before the cutover, it was automatically converted to a dollar balance at the same rate ($0.005 per credit). Your existing balance, monthly subscription, and billing all carry over. Nothing you need to do.

Are there volume discounts?

Yes. Contact hello@machine.dev for enterprise pricing and volume tiers.

Can I see my spend in the dashboard?

Yes. The dashboard shows per-job costs and daily/monthly totals. Cost Explorer groups spend by branch, commit, workflow, repository or runner. See Spend analysis for a walkthrough.

Do I pay for provisioning time?

No. You only pay for the time your job is actually running. Provisioning and teardown are free.

What happens if a spot instance is interrupted?

You’re billed up to the moment of interruption, then billing stops. A reclaimed runner can retry on on-demand before the job starts when fallback and interruption retry are enabled. Jobs already running need checkpointing and a workflow re-run. See Spot runners & savings.

Are storage charges separate from compute?

Yes. Compute is billed per minute based on the runner type. Storage (EBS) is billed prorated based on volume size, IOPS, and throughput. Defaults are minimal (~$0.006 for a 30-minute job at 100 GB / 6,000 IOPS / 250 MB/s).

Next steps

• Cost Optimization: strategies to lower your bill
• GPU Runners: pick the right GPU
• CPU Runners: pick the right CPU
• FAQ: billing, technical, and troubleshooting questions