Pricing
Pay for compute, not seats.
Choose a compute plan. Add AI credits only if you need them. Invite your whole team at no extra cost.
Cloud
from $20 / month
Ideal for individuals and teams
Claude Code, Codex, or your own keys. Tokens stay on your bill.
Total $20 / month
- Unlimited agents
- Up to 20 parallel sessions
- A full cloud computer per agent
- Connect your Claude Code or Codex subscription
- Any harness, any model, your own keys
- 100+ first party integrations
- Bring any MCP server
- Work with agents in Slack
- Jobs, schedules, and API triggers
- Apps with a database, storage, and a live URL
- Team memory and shared skills
- Approvals and audit logs on every action
- Spend caps and per-agent budgets
- Basic support
Enterprise
Custom pricing
Maximum security, compliance and control
- Everything in Cloud
- Deploy in your VPC
- Private inference — models never leave your perimeter
- Custom parallel scale — hundreds of sessions
- SSO and OIDC identity and access controls
- Fine grained roles, permissions and policies
- Org-wide secrets
- Detailed audit trails on every human and agent action
- Complete network and data residency control
- Uptime guarantees (SLAs)
- Dedicated account manager, forward deployed engineer and premium support
Compare plans.
| Cloudfrom $20 / month | Enterprisecustom pricing | |
|---|---|---|
| Eluu compute units (ECUs) | ||
| Included ECUs | 80 – 2,200 / month | Custom |
| AI credits (model tokens) | Optional add-on | Custom |
| Pooled across your team | ||
| Spend caps and per-agent budgets | ||
| Agent-hour cost reporting | ||
| Models & harnesses | ||
| Connect your Claude Code or Codex subscription | ||
| Bring your own keys (BYOK) | ||
| Any harness — no vendor lock-in | ||
| Private inference in your perimeter | ||
| Integrations & channels | ||
| First party integrations | 100+ | 100+ |
| Bring any MCP server | ||
| Work with agents in Slack | ||
| Kick off work from email or the API | ||
| Agents & workforce | ||
| Agents | Unlimited | Unlimited |
| Parallel sessions | Up to 20 | Hundreds |
| Full cloud computer per agent | ||
| Machine tiers | S / M / L | Custom |
| Team memory and shared skills | ||
| Jobs, schedules, triggers | ||
| Apps | ||
| Publish to a live subdomain | ||
| Custom domains | ||
| Database and object storage per app | ||
| Private / team / public sharing | ||
| Server apps with scale-to-zero | ||
| Security & governance | ||
| Per-agent, per-tool access control | ||
| Approval flows on sensitive actions | ||
| Secrets kept out of the model’s context | ||
| Audit logs | Standard | Detailed, org-wide |
| Secrets | User, project | User, project, org-wide |
| SSO and OIDC | ||
| Fine grained roles, permissions and policies | ||
| Deploy in your VPC | ||
| Data residency and network control | ||
| Support | ||
| Support level | Docs and basic support | Dedicated |
| Uptime guarantees (SLAs) | ||
| Forward deployed engineers | ||
Questions.
What is an Eluu compute unit?
An ECU is a normalized unit of the compute your agents run on — machine time, storage, apps. You get four ECUs per dollar, pooled across your whole team.
What are AI credits?
AI credits cover model tokens when Eluu is your LLM hub. If you bring your own Claude, Codex, or API keys, you do not need them — tokens stay on your existing bill.
Do I pay per seat?
No. Invite your whole team at no extra cost. You pay for what your agents actually do, not for headcount.
What does "agent hour" mean?
It is what an hour of an agent’s work costs you — compute plus tokens, divided by the hours it ran. It lets you weigh an agent against the job it does.
Can I run Eluu in my own cloud?
Yes, on Enterprise. We deploy the whole platform, including inference, inside your environment, so nothing leaves your perimeter.
