Hard spend caps for your OpenAI and Anthropic keys.
AI providers warn you when a budget is about to run out. They do not stop the charges. One stuck agent, one leaked key, one runaway batch job — and the bill is already in your inbox. aicostcap.dev gives you real hard stops, per-project alerts, and a dashboard that makes AI spend auditable.
Founding-member pricing locked for the first 100 signups. No card required.
Sound familiar?
- Your agent got stuck in a retry loop and burned $400 overnight.
- You handed an API key to a teammate and have no way to audit what they spent.
- You ship AI features in production and finance wants hard spend ceilings, not soft alerts.
- You run multiple side projects and refuse to share one billing account for all of them.
What aicostcap.dev does
Three primitives. Every other tool stops at soft alerts — this one actually stops the spend.
Hard spending caps
Set a ceiling per API key, project, or team. When the cap is hit we rotate the key to read-only — no more 3am bill-shock.
Slack + email alerts
50% / 80% / 100% thresholds fire in real time to Slack, email, or a webhook. Catch runaway agents before they catch your card.
Multi-provider dashboard
OpenAI, Anthropic, OpenRouter, Groq, Together. One dashboard, live spend, per-model breakdown, forecast-to-month-end.
Founding-member pricing
Locked in for the first 100 waitlist members. Public pricing will be higher at launch.
Solo
One engineer, three API keys, unlimited alerts.
- 3 API keys
- Slack + email alerts
- Per-project caps
- 30-day history
Team
Five seats, team-level budget, RBAC, audit log.
- 5 seats
- Unlimited API keys
- RBAC + audit log
- 90-day history
- Webhook + PagerDuty
Company
Unlimited seats, SSO, SLA, custom retention.
- Unlimited seats
- SSO (SAML / OIDC)
- SLA + priority support
- Custom data retention
- Quarterly reviews
Questions before you sign up
What does aicostcap.dev actually do?
It sits between your app and the LLM provider as a smart proxy — or watches the provider billing API for you — and shuts the tap off when a project hits its hard cap. You get per-key, per-project, and per-team ceilings that providers still do not give you natively.
Which providers are supported?
The MVP ships with OpenAI, Anthropic, and OpenRouter. Groq, Together AI, Google Gemini, and Mistral are next on the roadmap. If you use a provider that is not listed, reply to the waitlist confirmation email — it votes up priority.
Will I get billed when I join the waitlist?
No. The waitlist is free and gets you a founding-member discount when we launch ($29/mo solo, $99/mo team). You will only be charged if you choose to convert on launch day.
When does it launch?
The MVP ships within 4–6 weeks of hitting 100 waitlist signups. Founding members get a private Slack channel and a live walkthrough in week one.
Stop AI bills from surprising you.
One email when we launch. Founding-member pricing locked. Your work email only — no card, no spam.