LLMCap – A proxy that hard-stops LLM API calls when you hit a dollar cap
Source: Hacker News
Hard dollar caps
On every LLM call. When you hit $50, it stops. Not an alert — it stops. One line of code change. No surprise bills. Ever.
# your_app.py
from anthropic import Anthropic
client = Anthropic(api_key="sk-ant-...")
client = Anthropic(
api_key="sk-ant-...",
base_url="https://proxy.llmcap.io/anthropic"
)Works with every major provider
- Anthropic
- OpenAI
- Google Gemini
- Mistral
- Cohere
Setup in 5 minutes
How LLMCap works
- Providers supported: 5
- Avg added latency: <35 ms
- Requests blocked today: 18,742
- Uptime: 99.9%
Available everywhere you code.
Works in your workflow
VS Code Extension
Live spend in your status bar. Click to see today’s usage, burn rate, and blocked count — without leaving the editor.
Terminal CLI (PyPI)
Check spend, browse logs, and manage keys from the command line. Works on macOS, Linux, and Windows.
pip install llmcapWindows Tray App (Desktop)
System tray icon shows live spend. Right‑click for stats and quick actions. Always visible, never intrusive.
pip install "llmcap[tray]"Pick your plan
3‑day trial, no charge until it ends · Cancel anytime
Starter – $19 /mo (after 3‑day trial)
- 2 API keys
- All 5 providers
- Daily & monthly caps
- 30‑day audit log
- 1 user
- Email support
Credit card required for trial. Cancel before day 3 and you won’t be charged.
Questions
Does LLMCap ever see or store my API keys?
No. Your provider API key (e.g. sk-ant-...) is passed through the proxy header on each request and immediately discarded. LLMCap only stores your LLMCap proxy key, hashed with bcrypt. We never log your provider keys.
Does it work with streaming responses?
Yes — streaming is supported from day one. LLMCap passes SSE chunks through in real time. If the budget is exceeded mid‑stream, the connection is closed and a final 429 event is sent. The token that triggered the cap is not charged.
What exactly happens when the cap is hit?
The next incoming request is rejected with HTTP 429 before it reaches the provider. The token is never consumed, so you are never billed for it. Your app receives the same 429 response structure providers use for rate limiting, so existing error handling works as‑is.
Can I self‑host LLMCap?
Self‑hosting is on the roadmap. The proxy is open source (FastAPI + Redis). For now, the managed service at proxy.llmcap.io is the recommended path — it’s already deployed with <35 ms latency worldwide.