<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel><title>Citrusworks blog</title><link>https://citrusworks.io/blog</link><description>Practical writing on corporate brains, AI agents, evals and model choice, from the team at Citrusworks.</description><item><title>Five ways to use a voice profile</title><link>https://citrusworks.io/blog/five-ways-to-use-a-voice-profile</link><guid>https://citrusworks.io/blog/five-ways-to-use-a-voice-profile</guid><pubDate>2026-10-28</pubDate><description>A free MCP server that measures how you write, with five practical uses: checking AI drafts, writing across channels, diagnosing your style, ghostwriting and house style.</description></item><item><title>Jev vs OpenAI&#x27;s Decisions API: models that decide instead of write</title><link>https://citrusworks.io/blog/jev-vs-openai-decisions-api</link><guid>https://citrusworks.io/blog/jev-vs-openai-decisions-api</guid><pubDate>2026-10-12</pubDate><description>TypeSafe&#x27;s Jev and OpenAI&#x27;s Decisions API both return a typed decision with a probability instead of text. What each one is, what is still unproven, and what to do about it.</description></item><item><title>Pick the cheapest model that passes your evals</title><link>https://citrusworks.io/blog/cheapest-model-that-passes</link><guid>https://citrusworks.io/blog/cheapest-model-that-passes</guid><pubDate>2026-09-23</pubDate><description>Frontier, open or fine-tuned? Choose the cheapest, fastest model that passes your evals. Here is how to run that decision.</description></item><item><title>How to build a classifier that holds up in production</title><link>https://citrusworks.io/blog/how-to-build-a-classifier</link><guid>https://citrusworks.io/blog/how-to-build-a-classifier</guid><pubDate>2026-09-09</pubDate><description>Most AI work inside a company is classification. Here is how to choose between a prompt, a fine-tuned model and a classic model, and how to keep the result accurate.</description></item><item><title>When to train your own model, and how</title><link>https://citrusworks.io/blog/when-to-train-your-own-model</link><guid>https://citrusworks.io/blog/when-to-train-your-own-model</guid><pubDate>2026-08-26</pubDate><description>Fine-tune your own model when a small model is cheap and fast enough but not accurate enough, and you have labeled examples. Here is the decision and the process.</description></item><item><title>Every agent needs a judge</title><link>https://citrusworks.io/blog/every-agent-needs-a-judge</link><guid>https://citrusworks.io/blog/every-agent-needs-a-judge</guid><pubDate>2026-08-05</pubDate><description>An AI agent without evals is a demo. This is how to put a scoring gate in front of an agent so that only output that passes reaches a customer.</description></item><item><title>Build a personal AI OS that runs your week</title><link>https://citrusworks.io/blog/build-a-personal-ai-os</link><guid>https://citrusworks.io/blog/build-a-personal-ai-os</guid><pubDate>2026-07-29</pubDate><description>A personal AI operating system is a set of skills, memory and scheduled agents built around one person&#x27;s job. Here is what goes into one and how to start this week.</description></item><item><title>How to set up a team OS for AI agents</title><link>https://citrusworks.io/blog/how-to-set-up-a-team-os</link><guid>https://citrusworks.io/blog/how-to-set-up-a-team-os</guid><pubDate>2026-07-08</pubDate><description>A team OS is the shared set of skills, memory, agents and scheduled loops a team runs its work on. Here is how to set one up in five steps.</description></item><item><title>What a corporate brain is, and what it is not</title><link>https://citrusworks.io/blog/corporate-brain</link><guid>https://citrusworks.io/blog/corporate-brain</guid><pubDate>2026-06-17</pubDate><description>A corporate brain is one structured memory of what your company knows, readable by people and AI agents. What goes in it, and how to spot a fake.</description></item></channel></rss>
