My friend's Gemini bill jumped to $6K/month with zero visibility. So I built LLM Ops.
Hey IH! 👋
I just launched LLM Ops - a cost tracking and optimization platform for companies using AI APIs (Anthropic, OpenAI, Google).
Why I built this:
A founder friend recently told me his Gemini API bill jumped from a few hundred dollars to $6,000+ in one month. Zero visibility into what caused it. No breakdown by feature, no alerts, nothing. Just a massive surprise bill.
I spent time at AWS managing $2B+ in infrastructure spend, and I saw this pattern constantly with cloud costs. Now it's happening with LLMs - but moving 10x faster.
The problem:
Most startups building AI features have NO IDEA:
- What their actual per-request costs are
- Which features/teams are burning money
- When costs are spiking until the bill arrives
- That output tokens cost 3-10x more than input tokens
LLM pricing is deliberately confusing. A model advertised as "$10 per million tokens" actually costs $40+ for typical usage patterns.
What I built:
LLM Ops gives you:
✅ Real-time cost dashboards (by model, team, project, feature) for Anthropic, OpenAI and Google Gemini
✅ Anomaly detection & alerts (catch $6K bills before they happen)
✅ Department-level cost allocation
✅ Optimization recommendations
✅ 60-second proxy integration (no code changes)
What I'm working on:
- Automated optimization (smart routing + caching)
- Multi-cloud support
- Custom alerting rules
- Team budgets and spending limits
Would love feedback on:
- What LLM cost pain points are you facing?
- What features would make this a must-have for you?
- How much would you pay for 40-60% cost savings?
Check it out: https://cloudidr.com/llm-ops
Also built a free LLM pricing comparison page (60+ models): https://cloudidr.com/llm-pricing
Happy to answer any questions! 🚀
---
Background: Former AWS EC2 Capacity Products leader, managed $300M ARR. Built cloud infrastructure at scale for 6 years before starting CloudIDR.
