Self-hosting stays free. Temps Cloud hosts the control plane: the console, deploys, error tracking, analytics, backups and history for every server you connect. Your apps and traffic stay on your machines, billed by whoever you already pay for them.
Launch terms
14 days free · no card required
No per-seat pricing. No per-app pricing. No compute resale. PAYG starts off.
Two dials move between tiers: how many servers you connect, and how much data you send and keep. Seats and apps are never counted. Every plan also works with zero servers connected, for apps hosted elsewhere.
Self-hosted
For running Temps your way
Available now
Starter
For one or two servers
Coming soon
Team
Best for teamsFor teams running a fleet
Coming soon
Premium
Available now
Every Enterprise Edition feature for self-hosted Temps. One flat price per organisation: unlimited users and servers, nothing metered.
Get PremiumNeed an SLA, invoicing, an annual contract or custom terms? Talk to us
Extra server slots beyond the included count are $7 / server a month on either plan. Hosting is paid separately to your provider. Telemetry overage: Starter $0.50 / GB, Team $0.40 / GB, only after you enable a hard cap. Backup allowances use compressed GB-month, not source database size. Backup PAYG will remain off until its own hard-cap and reconciliation controls ship. Every backup receives integrity verification, and every plan includes a monthly restore-verification allowance shown above with no cap on run count; GB verified beyond that allowance bills pay-as-you-go at $0.01 / GB verified (1 GB min per run), only after you enable a hard cap. A compute hour is one vCPU-hour of actual query work—not a reserved CPU or an hour spent in the dashboard. Cached results and precomputed rollups are free.
Side project or just curious? Self-host Temps for free on any VPS. It is the same software, minus the hosted control plane. Install in 30 seconds
Starter includes 2 server slots and Team includes 10. Add any compatible server with an extra slot for $7 per month. You pay your hosting provider separately; hosting is not included in the slot price.
For example, 3 servers on Starter cost $36/month for Temps ($29 + one $7 slot), plus your hosting provider’s bill.
Connect a compatible VPS, bare-metal server or homelab server over SSH. Each server uses one slot from your plan or an extra slot you add. Hosting costs are paid separately to your provider.
Connect your supported cloud account and provision servers from the console. Each server uses one slot. Your provider bills your account directly for hosting; the Temps plan and extra slot price do not include that bill.
Transparent by design
Apps, databases and web traffic run on machines you own. The control plane deploys to them and watches them; it is never in the request path.
Run the join command on any VPS and it appears in your console. Starter includes 2 server slots, Team includes 10, then $7 per extra slot per month. You pay your provider separately for hosting.
Analytics, error history and backups are stored offsite and kept for as long as your plan says, so your servers never fill up and a lost box is never a lost history.
The fine print, plainly
Yes. Temps remains MIT / Apache 2.0 and free to run on your own infrastructure. Temps Cloud adds a managed telemetry datastore, fleet visibility, managed backups and usage billing without entering your application request path.
Each connected server uses one slot, whether it is a VPS, a bare-metal server, or a server in your own cloud account. Starter includes 2 slots and Team includes 10. Add extra slots to either plan for $7 per slot per month. Hosting is not included: you pay your hosting provider separately. Every connected server gets deployments, error tracking, analytics, request timing, session recordings and uptime checks.
Every new managed account starts with 14 days of Starter access. No card is required, and pay-as-you-go remains disabled unless you explicitly enable it and choose a hard monthly cap.
A compute hour is one vCPU doing query work for one hour, or the equivalent work spread across several cores. It measures the CPU actually consumed by analytics and OpenTelemetry queries—not reserved hardware or time spent viewing the dashboard. Cached results and precomputed rollups do not consume the allowance. Pay-as-you-go remains disabled until you explicitly set a hard monthly cap.
Backup allowances measure compressed data stored for one month, not the source database size. Every backup receives checksum, manifest and object-integrity verification. Every plan also includes a monthly restore-verification allowance against a disposable database — 10 GB on Starter, 20 GB on Team — with no cap on how many verification runs you trigger. Once the GB verified that month crosses the included allowance, additional GB verified bills pay-as-you-go at $0.01/GB (1 GB minimum per run), once you enable a hard spend cap.
Your apps, databases and web traffic stay on your servers. The hosted control plane never sits in the request path; it stores the console, analytics and error history, backups and configuration. If a server loses contact with the control plane, its apps keep serving.
Not to get started. The control plane is hosted, so apps deployed anywhere can send errors, analytics and request timing to it with zero servers connected. Connect your own servers when you want Temps to deploy and watch them.
No. Connect your own server over SSH, or create one through your own supported cloud account. You pay Temps for your plan and any extra server slots. You pay your hosting provider separately for the server, storage, bandwidth and other infrastructure charges.
Nothing you build locally becomes obsolete. Managed Cloud extends the same open-source Temps instance.
