The home of the Algorist

After Microsoft Copilot’s move from subscription requests to credit-based usage on 1 June 2026, the company has tried to lighten the bill shock by providing access to cheaper open weight models, such as Kimi K2.7 Code (available just over a one month later). Meanwhile, there have been plenty of even cheaper model options on another Microsoft enterprise platform hiding in plain sight: Microsoft Foundry. Read on to learn how to consume models you deploy yourself on MS Foundry alongside your regular options in both OpenCode and Copilot CLI.

I’ve been a member of the Ollama discord for a while now, and I’ve noticed that when people talk about Ollama Cloud they frequently compare it against OpenCode Go. As the Ollama free tier allowance has constrained over time (and my free tier GitHub Copilot went the same way), I finally bit the bullet and paid for AI using OpenCode’s subscription product for only $5 in the first month ($10 afterwards).

This post continues from AI llama-server on a phone with a detailed walkthrough of every step needed to flash LineageOS, set up Termux, build llama.cpp, and serve a small language model over your local network.

Yesterday I made an AI server run on my broken 6 year old phone. In the same week as GitHub moved to usage based pricing and a week after we learnt that a company accidentally spent $500 million on Claude AI in one month alone, I dug out a discard piece of technology to host an LLM.

Python recipe scraper exposed as an MCP tool