agent-guard

Prevent runaway AI agents from overspending

4 followers

Prevent runaway AI agents from overspending

4 followers

Contribute to adarshashokbaghel-code/agent-guard development by creating an acagent-guard is an open-source Node.js library that wraps OpenAI and Anthropic SDK calls to enforce hard spending limits, block calls once your budget is reached, and keep a local audit log—all without sending telemetry.count on GitHub.

Free

Launch tags:API•Open Source•Developer Tools

Launch Team

Wispr Flow: Dictation That Works EverywhereStop typing. Start speaking. 4x faster.

Promoted

Maker

📌

Hi everyone! 👋 I'm Adarsh, and I built agent-guard after noticing that most AI SDKs make it easy to monitor usage but don't provide a simple way to enforce a hard spending limit. If an autonomous workflow gets stuck in a loop, costs can grow much faster than expected. I wanted a lightweight solution that developers could add without changing the rest of their stack. With agent-guard you can: • Set a maximum budget in USD • Block API calls once the budget is reached • Keep a local audit log • Avoid sending telemetry to external services I'd really appreciate your feedback on the API design, CLI, and anything that could make it more useful for production AI applications. Thanks for checking it out!

Report

2mo ago

Reviews