Health AI Central

The problem

Health AI information is scattered: research literature, trial registries, regulatory databases, company announcements, and specialist reporting each tell part of a story, and nothing connects the story to the evidence. I wanted a publication that reads like specialist journalism and cites like a systematic review, and I wanted it to run every day without a newsroom.

The design

40+ agents do the daily editorial, research, and promotion work; the monitoring registry tracks 74 active workloads in all, 70 of them on a staggered overnight cron and 4 on demand. The model split follows one rule: Sonnet for judgment-heavy steps, Haiku for high-volume steps. Classifying a thousand headlines is volume; deciding what deserves 1,500 words is judgment.

The map below shows one slice, the editorial pipeline, the most visible of the agent families:

A night runs like this. Ingestion agents pull from news lanes covering 1,000+ companies, 25 RSS feeds, and primary sources (PubMed, ClinicalTrials.gov, FDA databases, NIH RePORTER), most without any model at all. A story listener (Sonnet) scores what came in; the content manager sets the day’s editorial brief; the medical researcher picks 10–12 stories and fetches each source’s full text; the technical writer (Sonnet) drafts news briefs of 400–700 words and deep dives of 1,200–2,000; the editor (Sonnet) reviews and calls the visual artist, which writes an image brief in Sonnet, renders it with Gemini 3 Pro Image, and then scores its own output.

Two other families work alongside the editorial pipeline. Research-intelligence agents screen the incoming data, link records across databases (a trial registration to its sponsor, its publications, and its regulatory record), and generate the intelligence products users come for: company profiles, clinical-area briefs, and classifications of why trials stopped. Back-office agents run the business itself: customer service and inbox triage, traffic analytics, SEO, security monitoring, advertiser operations, and the site's own help assistant. The same rules apply to every family: one gateway, per-agent keys, and a monitor that watches every job.

Hand-offs are files and database states, not vibes. The pipeline is a state machine with an explicit list of allowed transitions; an illegal transition raises an error instead of publishing a mistake. Every model call goes through one gateway with a per-agent key, so each agent’s spend and behavior are individually visible.

Evaluation and guardrails:

• The researcher pulls the primary records out of each source, a DOI, PMID, trial number, FDA clearance number, or NIH project number, and attaches them to the article so a reader can go past the story to the evidence.
• A source-fidelity reviewer (Sonnet) checks the draft against its sources; a duplicate-story guard checks it against everything already published.
• AI-generated images must score 80 or better in review; source images must pass a provenance check.
• At most two automated revisions, then a human hold.
• For research classification, the evaluation went further: a blind three-vendor judge panel decided disagreements by strict majority, and clinicians checked a subsample. Sealed result: F1 0.953.

And the guardrail that matters most: nothing publishes without me. The pipeline’s last automated act is an email. I approve, reject, or hold every article at Board Review, and my rejections are fed back into the next day’s story selection as negative examples. I reject about a quarter of what reaches me, which is the system working, not failing. In the early months it was half to 60 percent; adding narrow agents and strengthening each prompt brought it down.

The numbers

Figure	Value
Agents in the daily pipeline	40+ (74 active workloads in the monitor registry: 70 scheduled, 4 on demand)
Articles published since launch	500+
Drafted per day	~10–11
Published per day	5.3 → 6.9 (month over month, Oct 2026)
Quality-hold rate	34.7% → 14.5% after the contract change
My rejection rate at Board Review	~25% (50–60% in the early months)
Monthly model spend	≈ $200 via the gateway

What reaches me each morning

The pipeline's last automated act is an email. This is the one from November 6, 2026, exactly as it arrived, with addresses and system identifiers redacted: the image, the title and source, the result of every quality gate, the full draft, a verification report that lists each of the twelve claims with its verdict and the reviewer's note, the image reviewer's reasoning and score, the immutable review bundle with its fingerprints, and the three buttons. Anything else I write back is treated as feedback and goes to the writer for a revision.