Every price, on one page.
Per service, per month, billed per second. Monitoring, backups, image optimization, and the AI features are part of the price rather than an add-on you discover later.
Web services
Always-on APIs and applications. Priced per service, billed per second, with no cold starts on any plan — including free.
| Plan | CPU / memory | Regions | Autoscaling | Per month |
|---|---|---|---|---|
| FreeFree | shared · 512 MB | 1 | — | $0 |
| Starter | 0.5 vCPU · 1 GB | 3 | to 3 instances | $7 |
| Pro | up to 8 vCPU · 32 GB | all 15 | predictive, to 40 | $25 |
| Scale | dedicated capacity | custom | unlimited | Talk to us |
Static websites
The CDN is free on every plan. What you pay for is image optimization, the AI optimization agent, and bandwidth.
| Plan | Projects | Bandwidth | AI optimization | Per month |
|---|---|---|---|---|
| FreeFree | 2 | 100 GB | — | $0 |
| Developer | 10 | 1 TB | report only | $9 |
| Pro | unlimited | 5 TB | report + auto-fix PRs | $29 |
Background workers
Pay for worker time, not per message. The managed queue, scheduler, retries, and dead-letter storage are included.
| Unit | What it covers | Price | ||
|---|---|---|---|---|
| FreeFree | 1 worker · 50,000 jobs per month | $0 | ||
| Worker-hour | 0.5 vCPU · 512 MB, billed per second | $0.011 | ||
| Queue + scheduler | managed broker, no per-message fee | Included | ||
| Retries + dead letter | failed jobs kept 30 days | Free | ||
| Idle workers | scale to zero between bursts | $0 |
AI agents
Runtime is billed like any always-on service. Model tokens are passed through at provider cost — there is no markup on inference.
| Plan | Agents | Memory | Trace retention | Per month |
|---|---|---|---|---|
| FreeFree | 1 · 100 runs/day | short-term + conversation | 7 days | $0 |
| Team | unlimited | + long-term + shared | 90 days | $29 + usage |
| Enterprise | dedicated runtime, GPU options | all, per-region | unlimited | Talk to us |
Databases
PostgreSQL, MySQL, MariaDB, MongoDB, and Redis. The explorer, query editor, and assistant are included on every size — including free.
| Plan | Memory | Storage | Backup retention | Per month |
|---|---|---|---|---|
| Free — no expiryFree | 256 MB shared | 1 GB | — | $0 |
| Starter | 1 GB shared | 10 GB | 7 days | $7 |
| Standard | 4 GB dedicated | 50 GB | 14 days | $29 |
| Standard+ | 8 GB dedicated | 100 GB | 14 days | $59 |
| Performance | 16 GB dedicated | 250 GB | 30 days | $119 |
| Performance+ | 32 GB dedicated | 500 GB | 30 days | $229 |
| Custom / HA | 64 GB + | 1 TB + | 35 days | Talk to us |
Cloud storage
S3-compatible object storage. Tagging, OCR, captions, image variants, duplicate detection, search, and scanning are included rather than metered.
| Unit | What it covers | Price | ||
|---|---|---|---|---|
| FreeFree | 5 GB stored · 20 GB egress | $0 | ||
| Storage · hot | per GB per month | $0.020 | ||
| Storage · cold | after 30 days, same API | $0.008 | ||
| Storage · archive | restore in minutes | $0.002 | ||
| Egress | per GB, first 100 GB free monthly | $0.008 | ||
| Requests + search | no per-operation charge | Free | ||
| AI processing | tags, OCR, variants, dedupe, scanning | Included |
What a real stack costs
Nobody buys one service. These are three complete setups with every line itemised, so you can compare against what you pay now.
One API, one marketing site, one database.
Multi-region app, previews per pull request, real backups.
Plus provider token spend, shown per run and per customer.
Up to $5,000 in credit,
and engineers who do the move with you.
Bring your services across from any provider. We match your current annual spend in platform credit up to $5,000, and a Darwa engineer works alongside your team through cutover — including the parts nobody enjoys, like DNS, queues, and database replication.
Talk to the migration teamCredit is applied monthly against your usage and does not expire while you remain on a paid plan. Available to teams moving at least two production services.
The questions that decide it
No. Free web services do not sleep and free databases are not deleted after 30 days. Limits apply, a clock does not.
Per second for compute and per GB-month for storage, invoiced monthly. Resizing up or down takes effect immediately.
Bandwidth and storage overage are billed per unit and never suspend your service. You get an alert before each threshold.
No. Model usage is passed through at provider price, attributed per run and per customer so you can see where it went.
No. Logs, metrics, APM, health checks, and the AI diagnosis are included on every paid plan, and most of it on free.
Monthly by default. Annual prepayment earns two months, and Scale plans can commit for a discount and an SLA.