Flat pricing. Never a cut of your spend.
You pay your providers directly with your own keys. OpenPike is a flat monthly fee per workspace, whether your agents spend $50 or $50,000.
Free
For trying OpenPike with a side project or a single agent.
Start free- 3 virtual keys
- 10,000 requests a month
- Budgets and kill switches
- Fallbacks across your keys
- 3-day request logs
Team
Most popularFor engineering teams running agents in production.
Choose Team- 50 virtual keys
- 1,000,000 requests a month
- Loop detection
- Email and Slack alerts
- Spend by agent, developer and project
- 30-day request logs
Business
For companies with many teams, agents and budgets.
Choose Business- Unlimited virtual keys
- 10,000,000 requests a month
- Everything in Team
- Audit log
- Optional prompt logging
- 90-day request logs
- Priority support
Need SSO, SCIM or a contract with invoicing? Talk to us
Compare plans
| Feature | Free | Team | Business |
|---|---|---|---|
| Virtual keys | 3 | 50 | Unlimited |
| Requests per month | 10K | 1M | 10M |
| Token markup | None | None | None |
| Budgets and kill switches | Yes | Yes | Yes |
| Fallbacks | Yes | Yes | Yes |
| Loop detection | - | Yes | Yes |
| Slack alerts | - | Yes | Yes |
| Request log retention | 3 days | 30 days | 90 days |
| Audit log | - | - | Yes |
| Support | Community | Priority |
Questions
Do you mark up tokens?
No. You bring your own provider keys and pay providers directly at your own rates. OpenPike only charges the flat plan fee.
What happens when I hit my request limit?
You choose: OpenPike either slows requests or blocks them, and prompts you to upgrade. We never send a surprise overage bill.
What happens when a budget runs out?
Soft limits send an alert. Hard limits block requests for that key, agent or project until the period resets or you raise the limit.
Do you store my prompts?
No. Prompts and responses pass straight through. We keep request metadata such as model, tokens, latency and tags for spend reporting. Business plans can opt in to prompt logging.
What if my card fails?
We retry and email you. After 14 days the workspace moves to Free plan limits instead of shutting off, so your agents slow down rather than going dark.
Which providers work?
OpenAI, Anthropic, Google, DeepSeek, TypeSafe Jev, Groq, Mistral, xAI, Together, Fireworks, OpenRouter and more, with new providers added often.