
A developer combined Sakura Internet's rental server, Go, Cron, and Cloudflare's Clef decision model to score incoming mail. Each message is rated 0–100% for relevance and the score is prepended to the Subject line as tags like [AI:92%].
Summaries like this, in your inbox every morning.
The project started because spam was burying real inquiries from a company's web contact form, and the existing environment was a rental server that could not host an always-on process or a database. Rather than rebuild the stack, the developer leaned on what was already there: Cloudflare for DNS, Sakura Internet for mail, Go for a small binary, and Cron to run it every few minutes. The piece is a firsthand account of that constraint-driven design choice, not a product announcement.
What made it possible was a decision model rather than a text generator. Clef accepts text, JSON, images, or video as a state and returns probabilities for a defined question schema, so the developer could ask "is this an inquiry relevant to our company?" and get a score back. Workers AI serves @cf/cloudflare/clef, a 27B multimodal model, through its free tier, which the author notes keeps inference cost essentially zero at low mail volumes.
The operational result is modest but concrete: the score in the Subject line replaced manual scanning of the spam folder. The author argues the same pattern may lower the barrier to AI adoption for small setups, since Clef is specialized for judgment rather than generation and fits this kind of automated triage work well.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
ElevenLabs completed a $300 million employee tender at a $22 billion valuation, and co-founder Mati Staniszews…

On October 15, OpenAI announced the Rosalind Academic Plan, giving approved US university labs free access to…

A Qiita article compared automated red-teaming methods for LLM apps — GCG, PAIR, and TAP — using mock implemen…

Liquid AI released d1-ultra on October 15, 2026, an upgraded open-weights decision model that averaged 76.7 ac…

Circlia started a proof-of-concept for TASTE SIGNAL: visitors to Holland Village Cafe FUKUOKA taste cookies an…

In a webinar hosted by French Mathematical Society president Isabelle Gallagher, Fields Medal winner Hugo Dumi…
