manueljpconde gives mikeEU a one-command install and untangles its model settings

Two strands land together: a self-contained local stack any lawyer can spin up, and a clean split between public and self-hosted AI providers.

infrastructureintegration

The first half is a packaged local environment - frontend, backend, database, auth, and file storage all wired up to run on a laptop, with no hosted cloud credentials required. The pitch is explicitly lawyer-friendly: clone, run one command, kick the tyres.

The second half is a rethink of how the app handles AI models. Earlier work had quietly routed Microsoft's Azure and Foundry endpoints through the slot meant for plain OpenAI keys, which got messy. manueljpconde split the settings into two distinct tracks - public hosted models on one side, self-managed endpoints (Azure, Foundry, or a local model running on your own machine) on the other - each with its own configuration owned by the user or admin.

The author is upfront that this is scaffolding, not the finished story: connection checks, warnings for local URLs, and safer defaults for local models that fumble tool-calling are flagged as next up.

So what Worth a look for anyone evaluating Mike for a firm that wants to keep its AI stack - and its data - under its own roof.

View this fork on GitHub →

Spotted something wrong? Or know the PR text has fresher detail than the writeup above?

Commits in this thread

1 commit from manueljpconde/mikeEU, oldest first. Source extracted verbatim from the harvested git log.

SHA Subject Author Date
d39706fc Add Docker stack and managed model support Manuel Conde 2026-05-10 ↗ GitHub

Capture this thread into my fork

Download a single Markdown prompt that tells Claude how to port every commit above into your working tree — adapting paths and structure to match your repo. Run it via claude -p < capture-thread-252.md from inside the repo you want the changes in.

⬇ Download capture-thread-252.md