JevMade hello@JevMade.com
← Back to guides

JevMade field notes / Integration guide

A guide to preparing an AI assistant for incoming tasks

This guide explains how to use software that helps an AI assistant pick the right AI service level, executable tools, and saved instructions for a task before it starts working.

Original by JoacoMarcAgent workflows

Listen to this guide

JevMade’s plain-English explanation

0:00 /

AI narration

Credits and license

“jev-harness-router” by JoacoMarc. Read the original source.

This expanded guide is an AI-narrated adaptation of the source’s essential explanation, examples and caveats, not a word-for-word reading. The synthetic voice does not imitate the author or imply their endorsement.

License: MIT

MIT License

Copyright (c) 2026 Joaquín Marcoff

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.

Our summary

This project is a sorting tool for an AI assistant. When a user makes a request, the software decides how difficult the task is. It then chooses the right AI service level, executable tools, and saved instructions needed to complete the job without wasting time or money.

The setup uses Jev, an AI tool that chooses from a list of options instead of writing out an answer. It makes one fast request to pick the best setup from your saved lists. If the network takes too long, it falls back to simple keyword matching.

This guide is for people building AI software who want to separate task sorting from task execution. The tool requires a private access code to work. The actual financial costs and available AI services depend entirely on the external accounts the reader chooses to connect.

Key takeaways

  1. List your AI services from cheapest to most capable so the software can escalate when needed.
  2. Separate the step that decides how to handle a task from the step that actually does it.
  3. Test the sorting rules with the provided evaluation commands before changing how the software makes choices.

The success rates and response times reported in the guide happened in the author's specific tests on 181 examples. Actual performance and financial costs will vary based on the user's own setup and location.

Repository README · Source reviewed

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