Skip to content
ZYNXTECH

Index of workUpdated

Open-source tools for AI agents, money and people data.

Zynxtech builds and publishes developer tools, full-stack applications and Power BI reports. The two newest let coding agents benchmark models on your own hardware and see inside Power BI Desktop. Every project is MIT licensed and documented, with the source on GitHub.

pip install llmdyno
pip install pbidriver

Browse all 8 projects

01Current work

Both came out of the same problem. An AI coding agent can write the code, but it needs a way to check the result: which model actually finishes the task, and what the report looks like after the change.

01Developer toolv0.5.0 on PyPI

LLMDyno

Benchmark local and cloud LLMs as coding agents, on your own hardware.

LLMDyno runs real multi-file coding tasks through open-source agent harnesses in offline Docker sandboxes, grades them with hidden tests and blind LLM judges, and compares correctness, code quality, speed, VRAM and cost. Drive it from a web dashboard, from Python or from the command line.

agent harnesses, each in an offline sandbox
7agent harnesses, each in an offline sandbox
worst VRAM estimate error across 84 configurations
66 MiBworst VRAM estimate error across 84 configurations
API operations, each with a Python call and a CLI command
116API operations, each with a Python call and a CLI command
cloud providers, plus any OpenAI-compatible endpoint
15+cloud providers, plus any OpenAI-compatible endpoint
pip install llmdyno

02Developer toolv0.1.1 on PyPI

pbidriver

WebDriver for Power BI Desktop, so AI agents can see and drive the reports they edit.

Agents can already write DAX and edit PBIP files, but they cannot see the result. pbidriver gives them pixel-exact captures, the numbers behind every visual, slicers, menus, dialogs, DAX and Performance Analyzer timings, from a CLI that prints one JSON object per call.

commands, each printing one JSON object
25commands, each printing one JSON object
views it can capture, from the canvas to TMDL
5views it can capture, from the canvas to TMDL
dialog technologies it detects and answers
3dialog technologies it detects and answers
agents with setup guides, Claude Code to Cursor
4agents with setup guides, Claude Code to Cursor
pip install pbidriver

03Three threads

Eight projects, but they keep returning to the same three subjects.

A · 2 projects

Instruments for AI agents

A coding agent is only as useful as what it can check. LLMDyno measures which model and settings actually finish the work on your hardware. pbidriver lets an agent look at the Power BI report it just changed.

B · 2 projects

Money

Personal finance with real bookkeeping underneath, and a wealth management report that follows assets from the whole book down to a single adviser and client.

C · 4 projects

People

Appraisals, attrition, hiring and wellbeing. The records HR teams already keep, turned into workflows people can use and reports they can question.

04Recent releases

Full log
  1. pbidriverv0.1.1Documentation site and setup guides

  2. pbidriverv0.1.0First public release

  3. LLMDynov0.5.0The library can do everything the dashboard does

  4. LLMDynov0.3.1On PyPI

  5. LLMDynov0.3.0Python library

05Who builds this

Zynxtech is run by Jerome Eliezer, who writes the code, the tests and the documentation for every project on this site.

The work runs from Python services and React front ends to Docker sandboxes, Windows automation and DAX. The habits carry across all of it: run on the user's own machine where possible, publish the method with the numbers, and write the limits down next to the features.

About Zynxtech github.com/speak2jerome

Get in touch

Questions about a project, feedback, or an idea to work on together:

jeromeeliezer@zynxtech.in

Found a bug? An issue on the project's GitHub repository reaches the same person and helps the next user too.