JevMade hello@JevMade.com
← Back to guides

JevMade field notes / Integration guide

Guide to using Jev to check AI assistant actions

This guide explains how to use an AI tool called Jev to make quick decisions for your AI assistant. You will learn how to sort tasks and check for risky actions before the software carries them out.

Original by Sydney Runkle and Hunter LovellGuardrails

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Credits

“Building a Harness with Jev” by Sydney Runkle and Hunter Lovell. Read the original source.

This expanded guide is an AI-narrated adaptation prepared by JevMade. It expands the source’s essential ideas, examples and caveats in JevMade’s own words and is not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

Our summary

LangChain shows how to combine a standard AI assistant with Jev, a specialized AI tool that chooses from options rather than writing text. People use this setup to save time and money by letting Jev handle quick decisions while the main assistant focuses on writing and complex reasoning.

The guide provides examples of how this decision layer operates. In one setup, Jev reads an incoming request and selects the best AI model for the job. In another, it reviews the actions an AI assistant wants to take and flags potentially dangerous steps before they happen.

This approach is useful for developers building AI software who want to narrow how their assistant behaves. However, these examples do not prove the software will catch every dangerous action. Application permissions and human approval are still required to keep the system safe.

Key takeaways

  1. Use a standard AI assistant for writing, and use Jev for making quick, structured choices.
  2. You can set up rules so the software picks the right AI model for each specific task.
  3. A risk check can flag bad actions, but it does not replace strict system permissions.

The article provides examples of how to connect these tools together. It does not guarantee that the AI will successfully detect all risky behavior in a real project.

LangChain article · Source reviewed · Original published

Read the original guide Opens the author’s site in a new tab.