Publishing Studio
You bring the expertise. We handle the rest of it.
You don’t have to start from a blank page, wrestle with SEO, or build a structured article by hand. Ltrl drafts from work you’ve already done — a thread, a talk, a case, an explanation you give patients every week — and you approve what reflects your judgment.
𝕏Imported from X thread by @yourhandle
26 tweets · 2026-08-06
Chronic Inflammation: Mechanisms, Causes, and Evidence
A comprehensive review of inflammatory pathways, common triggers, and what the evidence supports.
Chronic inflammation represents one of the most pervasive pathophysiological mechanisms underlying modern disease. Research over the past two decades has established inflammatory processes as central contributors to conditions ranging from autoimmune disorders and cardiovascular disease to metabolic dysfunction and neurodegenerative conditions.
PMID Lookup
Induction of pro-inflammatory cytokines (IL-1 and IL-6)…
Conti P, Ronconi G, Caraffa A
J Biol Regul · 2020
PMID: 32171193
AI Suggested Citations (8)
“Research demonstrates that omega-6 and omega-3 fatty acids compete for incorporation into cell membranes.”
2016 · Nutrients
Simopoulos AP. An Increase in the Omega-6/Omega-3 Fatty Acid Ratio Increases the Risk for Obesity.
Basic AI chat versus the Ltrl Publishing Studio. Only one of them makes it compliant.
Both start from the same thread. Basic AI chat hands you a draft and stops there; the Studio keeps going, and everything after the draft is the part that makes it worth publishing under your name.
Basic AI
Write this thread up as an article
The Ltrl Publishing Studio
Magnesium Forms, Ranked by Absorption
Draft ready · claims marked, references resolved
- 1
It sounds like AI wrote it
Nothing you have written was read first, so it doesn’t sound like you — it sounds like everyone else. By the second paragraph it isn’t just a case, it’s a journey.
Your voice, mapped first
Modelled from what you've already written, refreshed every six months — it moves as you do.
- 2
The patient goes in as you typed them
Nothing is stripped, because nothing is looking. Pasting an identified case into a general-purpose tool is a disclosure you cannot take back.
HIPAA violationAnything patient-identifying is stripped
Names, record numbers, calendar dates, and other PHI are removed. How that works →
HIPAA Safe Harbor - 3
Claims stay buried in the prose
A statement of fact and the sentence around it are the same thing to it, so neither one gets checked.
0 claims checkedClaims are highlighted
Every factual claim is marked on its own, before a single reference is looked up.
12 claims identified - 4
Hallucinated references that look right
A citation gets written the way a sentence gets written: clean formatting, plausible authors, and a paper that often does not exist.
Citations are found and resolved
Anything you cited is kept exactly as you had it. Where a claim needs support, the Citation Engine looks one up and resolves it by its identifier.
ConsensusSemantic ScholarPubMed ✓DOI ✓ - 5
Keywords, repeated until it creaks
Ask a generic tool to make it good for SEO and it says your phrase eleven times and calls that structure. Nothing a search engine can map to a condition; nothing an answer engine can quote and attribute.
Structured for search and for AI answers
Headings and entities both engines can read — the step people call SEO, one of several.
- 6
A title, if you remember to ask
It might even be a good one — but it comes out of additional prompts rather than out of the finished piece, and nothing else comes with it. Eighteen prompts, then six dialogue boxes you fill in by hand.
Automated titles, slugs and descriptions
How the page appears in a result, written last, once the piece is settled. One click, and all of it is filled in.
- 7
It's finished when you close the tab
Nothing arrives anywhere. You copy it out, paste it somewhere, and every step above is one you had to know to ask for.
All six steps run on their own, then stop at you
You start it once and everything above happens end to end — no prompt, no second pass, nothing to remember to ask for. Nothing after this is automatic: the piece sits in your review queue until you approve it, edit it, or send it back.
Your Archive already works. A Ltrl Article makes it better.
A Ltrl Article is one continuous piece with titles, sections, references resolved against the literature, and an address of its own. Doing that by hand is an afternoon of restructuring, formatting and chasing references; here it’s the length of a coffee, and you keep the last word on every line.
Archived Post
Your thread, as you posted it
Magnesium: which forms actually absorb
Your name, MD @yourhandle · 4 Mar
Magnesium isn't one thing. What it's bound to decides how much of it you absorb.
1/5Oxide is the cheapest and the most common. Most of it never gets past the gut.
2/5Glycinate and citrate read better, though the trials measured them different ways.
3/5Threonate is studied for different reasons, and absorption isn't the question those trials asked.
4/5If you're comparing labels, the form is the part worth reading.
5/5
Ltrl Article
The same words, as one piece
Your name, MD · Perspective · 4 Mar
ltrl.com/providers/your-name/magnesium-forms-ranked-by-absorption
Magnesium Forms, Ranked by Absorption
Magnesium isn’t one thing: what it’s bound to decides how much of it you absorb. Oxide is the cheapest and the most common, and most of it never gets past the gut.1
Where the trials disagree
Glycinate and citrate read better, though the trials measured them different ways. Threonate is studied for different reasons, and absorption isn’t the question those trials asked.2
If you’re comparing labels, the form is the part worth reading.
References
- 1
Randomised crossover trial · serum response by form
PMID ✓ - 2
Systematic review · bioavailability measures compared
PMID ✓
Marked in yellow: the title, the sections and the resolved references the Studio added
Archived Post What’s already there | Ltrl Article What the Studio makes | |
|---|---|---|
| Your time | 0 min — it's already up | ~5–15 min to read it and say yes |
| Your words | Exactly as you posted them | The same words, assembled for you |
| How the posts sit | All on one page, in order | One continuous read, in sections |
| The title | Written for you from the thread | Written for the question people ask |
| Citations and references | Formatted as you posted them | Looked up, resolved and linked |
| SEO optimized | Partial | Yes |
| Where it's canonical | The original, wherever you posted it | Here — it's a new piece |
What you can publish
Gina Sterling, MD
Perspective
Why I stopped chasing TSH alone
TSH is a pituitary signal, not a tissue measurement. I kept seeing people whose number sat mid-range and whose symptoms did not.1
So I read free T4 and free T3 beside it. What that should change about treatment is still argued.2
4 citations · 3 questions answered
Case reports should be your best marketing
The people in your office already know how well you work — they watched you get it right. A case report shows the whole world. Look a clinician up on ZocDoc and you get a star rating; look them up here and you get what they have actually treated, written up under their own name.
Case report · published on Ltrl
Complete Symptom Resolution with Empirical Antimicrobial Treatment Despite Negative SIBO and Stool Testing in Functional Bacterial Overgrowth
Collin Dees, MPAS, PA-C
disorder of the small intestine · 3 citations
What it now answers
- Can SIBO be present even if breath testing is negative?
- Why treat empirically if diagnostic tests are negative?
- Why would a comprehensive stool test not show bacterial overgrowth in the small intestine?
- How does antibiotic history contribute to bacterial overgrowth?
- What is the rationale for delaying probiotics and prebiotics during antimicrobial treatment?
Clear gut symptoms, a negative breath test and a negative stool panel — Collin treated the clinical picture instead of the results. Under his name, on the condition it belongs to.
Read the full case reportPublishing that features your expertise, not just your star rating
Today a patient deciding between clinicians has star ratings and a directory row to go on. Publishing changes what there is to find.
What’s online now
- Star ratings
- Directory listings
- Referral networks
- A thin profile you didn't write
What you could put there
- The cases you have actually treated
- How you reason through a problem
- What you tell patients, in your words
- The conditions other practices send on
A star rating is a Google review score — a fair read on how visits feel, and only that. It can’t show whether you have seen their problem before. That gap is widest for the clinicians who take the cases other practices send on — you may not have the biggest front office, and you may still be the one who works out what is going on.
Nobody searches a star rating for these. They search the problem — and the page they find should be written by whoever has actually treated it.
Gregory House, MDDiagnostic Medicine · Princeton–Plainsboro★★★☆☆3.4 · 127 reviews★☆☆☆☆“Rude. Didn’t explain anything. Never looked up from the chart.”
Bedside manner
★★★★★“My son’s bloodwork kept coming back wrong. The lab called it contamination and ran it twice more. Dr. House asked whether both results could be right — two sets of DNA in one boy, from a twin he absorbed before he was born.”
Outcomes
Read the case reportSee what the other 125 sayEvery one of them is one of these two things
Gregory House, MDDiagnostic Medicine · Princeton–Plainsboro★★★☆☆3.4 · 127 reviews“Rude. Didn’t explain anything. Never looked up from the chart.”
Bedside manner
“My son’s bloodwork kept coming back wrong. The lab called it contamination and ran it twice more. Dr. House asked whether both results could be right — two sets of DNA in one boy, from a twin he absorbed before he was born.”
Outcomes
Read the case reportOne and two stars · 62 of them
- ★☆☆☆☆
“Never introduced himself. Told me my own doctor was an idiot, in front of my own doctor.”
- ★★☆☆☆
“Two hours in the waiting room. He was in with me for five minutes and spent three of them reading.”
- ★★☆☆☆
“Kept pushing about how much I drink. Wouldn't drop it. Made me feel like a suspect.”
Five stars · 51 of them
- ★★★★★
“Three months of pain in both knees and two scans read as normal. Everyone had stopped taking me seriously. He pulled the films up himself and said the thickening was the same on both legs, in the same segments, and stopped at my spine. Biopsy the next day. Fewer than three hundred people have ever been diagnosed with it.”
- ★★★★★
“Worked out in two days what three specialists missed in six months.”
- ★★★★★
“First seizure of my life at thirty-four, no history, scan came back clean. He asked what I had eaten and where. The cysts were on the film the whole time. Nobody had been looking, because nobody had asked.”
3.4 is the average of these two columns — the day he was short with someone and the day he was right, added together and divided. It is the only number a patient sees, and it describes neither one.
Illustrative. Reviews are written for this example; the clinical detail in them describes what each case involved and is not clinical guidance.
Published on a schedule, not on a Tuesday
Approving a piece and publishing it are not the same moment. Give an approved draft a date and the Studio takes it from there.
Pick a date, or don't
Publish now, or set it for later. A month's approvals can go out spaced across the month instead of all at once.
Search engines get told
The moment a scheduled piece goes live, its URL is submitted rather than waiting to be crawled. The clock starts the day it publishes, not the day it's noticed.
Nothing schedules itself
Only pieces you have already approved can be queued. Unapproved drafts stay drafts — there is no path from suggestion to published that skips you.
What every published page gets
Perspective
Magnesium Forms, Ranked by Absorption5
Which magnesium salts are absorbed best, what the bioavailability data actually shows, and how that changes what to take.
Your name, MD6
Your practice · verified provider
How absorption is measured1
Serum response tells you less than red blood cell magnesium3, which is why trials disagree.[1]2
Most head-to-head comparisons ran on serum, over hours, in small groups of healthy volunteers. The slower measures are the ones that separate the forms, and there are far fewer of them.
Related4
Read by search engines: Article · author · citations · medical entities · date7
1A clinical article structure
2Citation suggestions from your sources
3Medical entity linking
4Condition and biomarker relationships
5A search-ready title and summary
6Attribution to your provider profile
7Structured data for search engines
8A permanent URL that doesn't rot
Common Publishing Studio questions
Who actually writes the draft?
Ltrl does, from material you supply — your thread, your write-up, a PDF you upload. It isn’t writing about a topic from scratch; it’s restructuring your thinking into a form a reader and a search engine can both follow.
Are the citations real?
They’re looked up using leading medical search engines, not written, and there is an order to it. Anything you cited yourself is kept exactly as you had it. Where a claim needs support you didn’t supply, the Ltrl Citation Engine polls candidates from Consensus and Semantic Scholar — reading past the abstract rather than matching on a summary — and every suggestion is then resolved against PubMed by its identifier, so the reference on your page corresponds to a real indexed paper rather than to a plausible-looking string. The same engine binds the drugs and conditions in your piece to National Library of Medicine terminology and RxNorm, so what you wrote is tied to the entry an index already knows. The final call on whether a paper belongs against your claim is yours, which is why nothing publishes until you approve it.
How long does reviewing take?
One focused 15-minute session a month, covering two to four pieces.
What does de-identification mean here?
Case content is de-identified to the HIPAA Safe Harbor standard — no names, dates, record numbers, or details specific enough to single someone out. Calendar dates are removed outright, while relative timing is kept — day zero, week four — so the sequence of the case survives without a date anyone could match to a chart. A second pass then scans the draft for anything identifying that survived the first. Once it meets that standard it is no longer protected health information, which is the point: what gets published isn’t a patient’s record, it’s what you learned. You’re still the one who confirms it, because you’re the one who knows the case — and that confirmation is recorded, along with the de-identification run itself.
What happens to a draft I reject?
Nothing. It isn’t published, it isn’t indexed, and it doesn’t sit anywhere public. Rejecting costs you nothing and doesn’t need a reason — but if you say what was wrong with it, the same source can be run again. You never lose the thread it was drafted from.
You have the expertise. For 15 minutes a month, become the authority.
Bring one piece you’ve already written. The Publishing Studio drafts it, marks the claims and builds the page. Read it over, edit if needed — nothing publishes until you approve it.