OpenClaw pricing
OpenClaw Cost: Software, Hosting, API and Total Infrastructure Pricing
OpenClaw itself is free — MIT-licensed, no paid tier, no hosted service, no software fees. The cost of running OpenClaw comes from infrastructure (where the gateway runs), model access (which AI models it calls), and optional services (domains, backups, monitoring). The total depends on your architecture choices, not on OpenClaw licensing.
1. OpenClaw Software Cost
| Component | Cost | Source |
|---|---|---|
| OpenClaw agent | Free (MIT license) | OpenClaw Foundation |
| Updates | Free (open-source releases) | github.com/openclaw/openclaw |
| Support | Community (no paid support tier) | OpenClaw Foundation (501(c)(3)) |
The OpenClaw Foundation is an independent 501(c)(3) nonprofit. It accepts donations but does not charge for the software. There is no commercial tier, no per-seat pricing, and no hosted service.
2. Hosting and Hardware Cost
Where the gateway runs determines the infrastructure bill. These are documented provider prices, not AIG-tested recommendations.
| Option | Monthly cost | Notes |
|---|---|---|
| Existing computer | $0 (electricity only) | No new hardware. You own uptime and connectivity. |
| VPS (2 GB RAM) | $10-12/mo | Vultr $10, DigitalOcean $12, Linode $12, Fly.io ~$11. Documented provider pricing; not AIG-tested for OpenClaw. |
| VPS (4 GB RAM) | $12-24/mo | Hostinger KVM 1 $6.49 promo / $11.99 renewal (4 GB). DigitalOcean $24 (4 GB). |
| Mac mini M6 (16 GB) | $899 upfront + electricity | One-time purchase. No monthly hosting bill. Apple Silicon. |
| Mac mini M5 Pro (24-64 GB) | $1,699+ upfront + electricity | For local model inference alongside the gateway. |
VPS prices checked September 17, 2026 from official provider pages. Mac mini prices from apple.com/mac-mini/specs. Promotional prices noted where applicable.
3. Model and API Cost
OpenClaw routes requests to model providers. The API cost depends on which provider and model you use and how much you use it. For many deployments, API cost will exceed hosting cost.
| Provider | Example model | Input | Output | Source |
|---|---|---|---|---|
| OpenAI | Current API models | Varies by model | Varies by model | See official pricing |
| Anthropic | Current Claude models | Varies by model | Varies by model | See official pricing |
| Ollama (local) | Any supported model | $0 | $0 | Compute is in hardware |
API pricing changes frequently. Use the official provider pricing pages linked below for current model rates before budgeting.
4. Local Model Cost
Running models locally via Ollama eliminates per-token API costs but shifts the expense to hardware and electricity:
- Hardware: Mac mini M6 from $899, M5 Pro from $1,699. Memory configuration depends on model size.
- Electricity: monthly electricity cost depends on the hardware configuration, workload, runtime, and local electricity rate.
- Amortization: A $899 Mac mini over 3 years is ~$25/mo before electricity — comparable to a VPS but with local model inference included.
- No per-token cost: Once the hardware is purchased, local inference has no usage-based billing.
5. Optional Costs
| Service | Cost | Notes |
|---|---|---|
| Domain name | ~$10-15/year | Optional — only needed for custom remote access |
| Tailscale | Free (personal use) | Documented integration for secure remote access |
| SSL certificate | Free (Let's Encrypt) | Required if exposing gateway publicly |
| External backup | ~$0.02/GB/mo | S3-compatible storage. OpenClaw state is small. |
| Monitoring | Free tier available | Recommended for always-on deployments |
Cost Scenarios
These are calculated examples based on documented pricing, not measured results. Your actual cost depends on workload, provider, model selection, and usage patterns.
| Scenario | Monthly estimate | Assumptions |
|---|---|---|
| Existing Mac + Ollama | Electricity cost | No new hardware. Electricity only. No API cost. Evidence: CALCULATED |
| VPS + light cloud API | VPS + API usage | VPS price plus actual model API usage. Calculate the API portion from your provider, model, and token volume. Evidence: CALCULATED from documented pricing |
| VPS + moderate cloud API | VPS + API usage | VPS price plus actual model API usage. Active tool use can change token volume substantially by workload. Evidence: CALCULATED from documented pricing |
| VPS + heavy API usage | Workload-dependent | API cost dominates hosting cost at high usage. Evidence: CALCULATED from documented pricing |
| Mac mini M6 + Ollama | ~$25 amortized + electricity | $899 over 36 months ≈ $25/mo. No API cost. Electricity varies. Evidence: CALCULATED from Apple pricing |
Model your own OpenClaw hosting cost
Starts on the Mac mini M6 scenario from the table: $899 of hardware over 36 months. Switch the hosting model to VPS or GPU cloud to price the other scenarios. Model API usage is not included — add it to the extras field from your own provider and token volume.
Your next step
At this monthly equivalent, owned hardware is the cheaper option, so the next decision is how to specify and operate it.
Plan the local AI server behind this cost →Open the full hosting cost calculator → for the formula and what the estimate deliberately leaves out.
Sources
- OpenClaw — Documentation— MIT license, no paid tier, no hosted service
- OpenClaw Foundation— 501(c)(3) nonprofit, MIT license
- Apple — Mac mini Specifications— Current pricing and configurations (September 2026)
- OpenAI — API Pricing— Official current model and token pricing
- Anthropic — Pricing— Official current model and token pricing
Agent Infrastructure Recommendation Kit
Need help modeling your OpenClaw infrastructure cost?
The Recommendation Kit includes cost-comparison worksheets for VPS, local hardware, and hybrid architectures — helping you document assumptions, compare options, and make an evidence-based infrastructure decision.
Get the Recommendation Kit — $19