Know what you're
spending on AI
Privacy-first cost tracking proxy for AI APIs. Set budgets, track spend across models, never get surprised. All data stays on your machine.
Requests today: 47
Top model: claude-sonnet-4 ($3.81)
Status: ● healthy — 42% of daily limit
Everything you need to control AI spend
Privacy-first
Runs entirely on your machine. No cloud, no telemetry, no tracking. Your prompts and costs stay yours.
Multi-provider tracking
One proxy for OpenAI, Anthropic, and more. Unified cost tracking across all your AI usage.
Budget Controls
Pre-flight checks, soft/hard limits, model suggestions. Let your agents self-regulate their spend.
Per-request costs
Know exactly what each API call costs. Track by model, by day, by whatever you need.
Hard limits
Set daily or monthly limits. When you hit them, requests stop. No surprises on your bill.
Drop-in proxy
Point your SDK at localhost:8787. No code changes needed. Works with any OpenAI-compatible client.
Up and running in 60 seconds
Clone & install
Clone the repo and install globally.
git clone https://github.com/n1byn1kt/agentcost
cd agentcost && npm install && npm install -g .
Start the proxy
Run the agent. It starts a local proxy on port 8787.
agentcost agent
Point your apps
Set your base URL to localhost:8787. That's it.
export OPENAI_BASE_URL=http://localhost:8787/openai
Take control of your AI spend
Open source, self-hosted, and built for developers who want visibility into every API call.