AgentWarden
Open source proxy that cuts repeated context for long running agents.
An open source proxy that sits in front of OpenAI compatible endpoints and cuts what long running agents burn on repeated context. Built in five days during OpenAI Build Week.
Agents re-send the same context on every step, and you pay for it every time. AgentWarden sits between the agent and the endpoint, spots what is being repeated, and strips it before the call goes out. Your API key never leaves your own process. It ships as a pip install with a local dashboard, and it is the only thing here whose source anyone can read, which is why it is the first link I send.
Python, FastAPI
![EthCC[5], Paris](/_next/image?url=%2Fphotos%2Fstrip%2F01-IMG_6659.jpg&w=3840&q=75)




