Pick a plan that fits your team. Credits, concurrency, and AI agents scale with you.
3,000 · valid for 30 days
Build your first AI agent, connect your tools, and watch a run end to end.
Plan $0 · member packages $20–$200 each
More credits and concurrency for teams running AI agents every day.
Plan $160 · member packages $20–$200 each
Room for a team of AI employees: high credit volume and 10 agents running at once.
Custom credit volume
Advanced security, dedicated support, and custom limits for your organization.
One package per member. Any overage uses pay-as-you-go credits.
20,400 credits
400 bonus included
52,600 credits
2,600 bonus included
108,700 credits
8,700 bonus included
222,200 credits
22,200 bonus included
On paid plans, top up anytime at 1,000 credits per $1. Set up auto-recharge so your agents never stop - when your balance drops below a threshold, credits are purchased automatically.
| Features | Free | Pro | Team | Enterprise |
|---|---|---|---|---|
| Credits & runs | ||||
Credits per member Credits included in each selected monthly member package | 3,000 · valid for 30 days | 20,400–222,200 credits | 20,400–222,200 credits | Custom |
Concurrent runs Number of agents that can run concurrently | 1 | 2 | 10 | Custom |
Extra concurrency Add concurrent run slots on top of your plan whenever you need them | — | — | ||
Org-shared agents Agents published to everyone in your workspace | 7 | 7 | 7 | 7 |
Private agents Agents only you can see | Unlimited | Unlimited | Unlimited | Unlimited |
Credit top-up Purchase additional credits anytime at $1 per 1,000 credits | — | |||
Auto-recharge Automatically purchase credits when balance falls below a threshold | — | |||
| Agentic automation & triggers | ||||
Scheduled runs Run agents on a recurring schedule for daily briefs, monitoring, and digests | ||||
Email triggers Start a run when a new email arrives or a label is applied | ||||
Calendar & meeting triggers Start a run when a calendar event changes or a meeting transcript is ready | ||||
Webhook triggers Start a run from any external system with a signed webhook | — | — | ||
Reusable workflows & skills Save a task once, then run it from any agent or automation | ||||
| Models & generation | ||||
Models Models you can choose for a run | 3 models | All models | All models | All models |
Bring your own LLM Use your own model provider API keys | — | |||
Org model policy Admins set the default model and provider for everyone in the workspace | ||||
Image generation Generate images inside a run | ||||
Video generation Generate video inside a run | — | |||
Presentation & website generation Turn a run into a finished deck or website you can share | ||||
| Connectors & channels | ||||
Built-in connectors Official connectors maintained by Okou, including X, Firecrawl, GitHub, Slack, and Notion | All connectors | All connectors | All connectors | All connectors |
Custom connectors Add your own internal APIs as connectors, with secrets stored outside agent code | ||||
Web services Web search, page scraping, maps and places, weather, market data, and people search, with no account or API key of your own | ||||
Channels Talk to your agents from the web app, Slack, Telegram, or by phone and SMS | ||||
Email Agents prepare Gmail drafts in your account for you to review and send | ||||
Voice input Hands-free speech-to-text. Limits apply per member and reset daily. | 10 lifetime requests · 10 min/day | 300 requests/day · 200 min/day | 500 requests/day · 500 min/day | Custom |
Computer use Let an agent operate applications on a desktop you connect | ||||
| Delivery | ||||
Hosted sites Publish an agent's HTML output to a public URL | ||||
File exports Download run outputs and generated files | ||||
| Security & governance | ||||
Sandboxed execution Firecracker microVMs with hardware-level KVM isolation | ||||
Full audit trail Agent HTTP/HTTPS traffic logged with SHA-256 integrity per run | ||||
No credential exposure Secrets injected at network layer, never visible to agent code | ||||
Per-permission access control Approve each connector permission separately instead of handing over a whole token | ||||
Per-member credentials Every member connects their own accounts, and agents act with the caller's access | ||||
Time-boxed access Grant access that expires on its own | ||||
| Collaboration | ||||
Team members Number of people in your workspace | Unlimited | Unlimited | Unlimited | Unlimited |
Member usage dashboard Admins see each member's credit usage rolled up by agent | ||||
| Support | ||||
Community support Access to Discord community and documentation | ||||
Email support Direct email support from the Okou team | — | |||
Priority support Dedicated support with faster response times | — | — | ||
Credits are consumed when your agents use AI models. Different models consume credits at different rates. For example, a simple task might use a few credits, while a complex multi-step workflow uses more.
Yes. Every new workspace starts on Free with 3,000 credits valid for 30 days, one concurrent run, unlimited agents, and all connectors. Free runs on three included models. Upgrade to Pro for the full model list, your own provider keys, credit top-up, and auto-recharge.
Yes. Increasing a member package is prorated and takes effect after payment. Decreasing a package takes effect on the next billing date. Before you confirm, we show the charge and effective date.
On Free, upgrade or use a redemption code. On paid plans, buy credits at $1 per 1,000 credits, turn on auto-recharge, or choose a larger member package.
Monthly member-package credits expire at the end of the billing period and do not roll over. Credits purchased separately never expire. Credits that expire sooner are used first.
Existing credits keep their original expiration dates. Increasing a member package adds prorated purchased and bonus credits after payment. Decreases start on the next billing date.
Pro, Team, and Enterprise support your own LLM API keys (Anthropic, etc.). Free does not. Using your own keys does not consume Okou credits for model usage.
Every agent run executes in an isolated Firecracker microVM with hardware-level KVM isolation. Credentials are injected at the network layer and never exposed to agent code. All agent HTTP/HTTPS traffic is logged with SHA-256 integrity per run.
Contact us to discuss volume pricing, annual billing discounts, and custom arrangements for your team.