Updates

How LEGH.org changes, and why — infrastructure, privacy, and companion updates as they happen.

We Corrected Statistics in Our Stigma Article

We traced the statistics in our article on stigma to their original sources and corrected the ones that did not match.

What changed

  • Youth suicide. The article now says what the figure measures: the suicide death rate among Black youth ages 10 to 17, over the years 2007 to 2020, with the numbers and the report it comes from.
  • Suicidal thoughts. The article now uses the national survey figure for Black young adults. A smaller online study that found a higher rate is described as a smaller, non-national sample.
  • Treatment. The article now uses the most recent national survey, and compares treatment among adults with a mental illness, in place of an older figure that measured a different group.
  • Sources are now listed at the end of the article.
  • Oji Echo and Hope Echo received the same corrections in the scrolls they draw from.

What did not change

  • The article’s purpose: naming stigma and pointing toward support.
  • LEGH’s companions, articles, and education are free, and they stay free.

If you see a figure anywhere on the site that does not match its source, tell us.

Read More →

We Added a Page on How to Read Our Research

Studies do not always agree, and research keeps changing. So we added one page that explains how we use sources, what to do when two credible sources point in different directions, and how we correct mistakes. You can find it at How to Read Our Research. It is linked from the footer and from every article.

What changed

  • A new page, How to Read Our Research.
  • Our grief article now describes the follow-up analysis with Black participants more fully, including where the results were less clear, and describes one brain-imaging finding as coming from a small study.
  • Our article on the power of words now names the brain regions in social-pain research more broadly, and describes self-talk and brain change as encouraging but less settled.
  • Our article on racial stress and inheritance now says that whether stress effects pass to children is still being studied and the findings are mixed.
  • Our articles on purpose, perseverance, and connection now describe findings about stress and the brain as associations and tendencies.
  • Oji Echo and Hope Echo received matching updates in the scrolls they draw from.

What did not change

  • LEGH’s companions, articles, and education are free, and they stay free.

If you see a figure anywhere on the site that does not match its source, tell us.

Read More →

We Made Our Wording About Research and Results More Careful

We want every sentence about research to say what the research actually supports. This round, we went through our articles, wellness notes, Natural Mystic modules, and Know Your Roots pages, and the knowledge our companions draw on, looking for places that stated a finding or an effect more strongly than the evidence allows.

What changed

  • Sentences about research findings now say what the research suggests, supports, or points to, instead of describing it as settled proof.
  • Sentences about what music, community, or a practice does for a person now describe what it can do or how it can help, instead of promising a result.
  • Words that borrowed clinical authority were replaced with plainer language, or with the name of the field that makes the point.
  • Oji Echo and Hope Echo received the same wording updates in the scrolls they draw from, so what they say matches what the pages say.

What did not change

  • The research and the named sources behind our articles are the same.
  • Where we quote a person or a study in its own words, we left those words as they were.
  • LEGH’s companions, articles, and education are free, and they stay free.

If you see a line anywhere on the site that says more than the evidence supports, tell us.

Read More →

We Made Our Wording About Brain Research and Pain More Careful

Several of our pages described a brain-imaging finding more strongly than the research supports. We have corrected them.

What changed

  • Pages on respect, words, dignity, shame, envy, and belonging now say that social rejection and exclusion engage some of the same brain regions as physical pain, that those regions do other work too, and that the patterns can be told apart. Overlap, not sameness.
  • What the research does support stays in the text: being excluded or disrespected is a real stressor the body registers.
  • Oji Echo and Hope Echo received the same corrections in the scrolls they draw from.

What did not change

  • The articles’ purpose and the named researchers behind them.
  • LEGH’s companions, articles, and education are free, and they stay free.

If you see a line anywhere on the site that says more than the research supports, tell us.

Read More →

We Corrected Figures in Our Grief Article

We checked the numbers in our article on prolonged grief against the original studies and corrected what did not match.

What changed

  • Treatment results now describe what the clinical trials found, and who was in them, in place of a single percentage.
  • How common prolonged grief is now cites the review of fourteen studies the figure comes from.
  • Diagnostic timelines now name both the DSM-5-TR and the ICD-11, since they differ.
  • Brain imaging findings are described as early, because they rest on small studies.
  • Statistics on losing family members now say what the national study measured, and that it measured deaths, not grief.
  • Depression alongside grief now gives the range reported in earlier studies, with its source.
  • Faith and community are described as a source of support many people find, since the research on how and for whom is mixed.
  • A Know Your Roots page and a wellness note received small wording corrections.
  • Oji Echo and Hope Echo received the same corrections in the scrolls they draw from.

What did not change

  • The article’s purpose: naming prolonged grief, and pointing toward support.
  • LEGH’s companions, articles, and education are free, and they stay free.

If you see a figure anywhere on the site that does not match its source, tell us.

Read More →

We Simplified Our Daily Quotes

We want our use of other people’s words to be simple and clear. This round, we made our daily quotes easier to stand behind.

What changed

  • Sources we kept. Our daily quotes now come from historical speeches, books, scripture, and hymns in the public domain, each with a named source.

  • Quotations from songs, interviews, films, and television were removed from the daily quote feature. Because of this, the Music Wisdom page has been retired. If you saved its address, it now sends you to the main quotes page.

    Read More →

We Made How We Describe LEGH More Exact

We want every sentence about LEGH to be one we can stand behind. This round, we made the way we describe LEGH more exact.

What changed

  • The footer on every page, our promises page, and a few article and bio footers now describe LEGH in more exact words.

What did not change

  • LEGH’s companions, articles, and education are free, and they stay free.
  • Every promise on our promises page still stands, including the one about keeping LEGH separate from anything built for profit.

If you see a line anywhere on the site that says more than is true, tell us.

Read More →

We Reworded Phrases, Updated Quote Commentary, and Gave Three Pages Matching Web Addresses

We keep checking our pages for statements that describe why something happened without a source behind it. This round found two more in our pages and a few more in the commentary beside our daily quotes, and we fixed three web addresses that still carried old wording.

Two lines reworded

  • Natural Mystic, Africa Unite. A line stated a purpose behind what our ancestors lived through without a source. It now describes the effect that is documented: separations, lost records, and silences that left gaps in many family histories.
  • Culture, Identity, and Behavior. A line made a claim about the intended purpose of commercially packaged content. It now says only that such content rarely connects people to elders and community.

Daily quote commentary updated

We reviewed the explanations that appear beside our daily quotes. A few lines described why something happened or what a system intended without a source, and they now say only what is documented. We also made our own wording about pain plainer. The quotes themselves, in their authors’ own words, were not changed.

Read More →

We Added Evidence Notes to Three Pages and to Our Companions

Some research and some history is widely repeated but still debated or undocumented. We reviewed where our pages and our companions state such things as settled, and added a plain note where the evidence is mixed.

Pages

  • The Weight Has a History. Claude Steele’s idea of stereotype threat is now described as a debated finding. Many studies found it, the size of the effect varies, and some replications did not find it. The article no longer says Black students “score lower” as a settled fact.
  • She Went Back: Harriet Tubman. Several lines were stated more firmly than the record supports, and the page now says what is documented. Narcolepsy is a proposed diagnosis, not a confirmed one; what is documented is her head injury and her sudden sleep spells. The very large reward figures often repeated about her are not documented. She served as a scout and guide on the Combahee River Raid, which Colonel James Montgomery commanded, and some accounts call her the first woman to lead an armed operation, a claim we now hold loosely. She led a small team of Black scouts, and accounts differ on whether there were eight or nine. We no longer call her the leader of an entire intelligence network.
  • Understanding PTSD. Dr. Joy DeGruy’s term, post-traumatic slave syndrome, is now marked as a proposed framework and not a clinical diagnosis.

Our companions

Oji Echo and Hope Echo now say the same things, and a few more, when asked:

Read More →

We Checked Our Claims About Intent and Added a New Article

We reviewed our articles for one kind of claim: statements about why something happened, such as “designed to divide.” We checked each against primary or reputable sources. Where a source documents what a law or practice did, we say that. Where we could not find a source for the intent, we removed the motive.

What changed

  • Intent language. Phrases like “designed to,” “deliberate strategy,” and “deliberately destroyed” came out where we had no named source for the intent. Articles affected: Unity, Internalized Racism, Systemic Inequities, Community Violence, Hyperindividualism, The Village, Schadenfreude, One Love, and When We Eat.
  • What stayed. Documented intent stays, with its source named. The Senate’s Church Committee found in 1976 that the FBI worked to disrupt Black organizations under COINTELPRO. We say that, and we cite it.
  • Slavery and culture. We now state what the record shows: the 1740 Negro Act banned drums and horns, slave codes barred teaching enslaved people to read, and families were separated by sale. We removed claims we could not verify, including a general ban on African languages and a story about forced ethnic mixing.
  • Series name. “The Wound and the Medicine” is now “Turning Inward, Turning Together.” Several pages now say “protection” or “wellness” where they said “medicine.”
  • Human nature. Where pages said people were “designed” for community, they now say we evolved for it.

New article

Can Racism Change? What the Evidence Actually Says looks at what research shows about how people sort each other, and about what changes it.

Read More →

We Corrected Two Know Your Roots Pages

We checked two Know Your Roots pages against the sources, and we found errors. We fixed them today. Here is what changed.

Miriam Elizabeth Benjamin

  • How her chair worked. The page said it ran on an electrical circuit. Her patent describes a purely mechanical design: a gong knob, rods, and a crank that turned a red signal ball.
  • What her patent shows. The page said she invented the call button. She was one of many inventors working on signaling devices in her era. Her patent shows she was working on the idea early and put it on the public record.
  • Samuel Lee. The page said she tracked Lee down. Lee died in 1895. The 1906 award was paid after his death, and the page now says so. The amount now reads as the 1906 Boston Daily Globe reported it, and the page says she brought the bill before Congress nine times.
  • Her shoe sole. The page said it delivered medication. Her patent describes a pouch of powdered material, such as sulfur, meant to keep feet warm and help circulation.
  • A few smaller fixes. The pinking device is now marked as reported, not confirmed. The Civil War was about five months old at her birth, not four. We removed a street address we could not verify, and a named speaker for a quotation about her mother.

Bessie Blount Griffin

  • Held loosely, as sources differ. Her parents’ names, the spelling of her nursing hospital, and her exact degrees now say so.
  • Marked as accounts. Her work with veterans, her career dates, and the claim that she was the first Black woman trained at Scotland Yard now say they are widely reported but not confirmed.
  • Removed. A line saying the FBI turned away her application, which we could not source. We also removed a detail about building her feeder in her kitchen over ten months, and a stated reason for the feeder’s design.
  • Her patent. The page now gives the number, U.S. 2,550,554, and says the patent covers a neck-worn support for drinking, not the Invalid Feeder.
  • The title and opening. They no longer state as fact an account we could not confirm.

What isn’t finished

We have not yet confirmed the exact payee and amount in the Lee award. We will update the page if we do.

Read More →

We Clarified How LEGH Names Its Founder and Its Separate Work

We made a few small changes to how this site describes who is behind LEGH. Here is what changed.

  • The founder’s name. The About page and the articles now name the founder as David M. Gravely, also known as IKABU. Before, the site used only the name IKABU in most places.
  • The footer. The copyright line now names the founder, “© 2026 David M. Gravely,” and no longer names a company or a technology.
  • Our commitment to keep LEGH separate from anything built for profit. The promise has not changed. The page now says “LEGH’s founder also develops commercial technology separately” and no longer names a company. Nothing built for profit will ever be paid for by charging LEGH’s community, selling their information, or using their conversations.

We would rather tell you what changed than leave you to find it.

Read More →

We Named Everyone Who Touches Your Words — and Wrote Down What We Will Not Do

On September 15, we told you Oji Echo’s conversations travel through outside servers. Today we’re giving you the details we left out, and a few promises in writing.

We named the outside services

Three kinds of outside parties are in the path of your words:

  • DigitalOcean hosts LEGH’s own companion server, where saved conversations live until they’re deleted after 55 days.
  • OpenRouter is the routing service that carries your message to the AI model and the reply back.
  • The model provider. OpenRouter chooses which company runs the model for each message. We don’t pick. The ones eligible today are DekaLLM, NextBit, DeepInfra, CoreWeave, Venice, Parasail, NovitaAI, SiliconFlow, io.net, and Google Vertex.

They are different companies in different places, so your message may be processed outside the United States. DekaLLM, for example, runs its servers in Indonesia, and io.net describes itself as a decentralized network of computers.

Read More →

The Oji Echo Chat Page No Longer Loads Anything From Google

Earlier today we told you the Oji Echo chat page loaded its typeface from Google Fonts, which meant Google could see your device’s connection address when the page opened. We said we were working toward fixing that.

It’s fixed. The chat page now uses the same built-in fonts as the rest of LEGH.org: the typeface your own device already has. Nothing is downloaded from Google when you open it.

The text looks a little different from before. The words, the conversation, and how your messages are handled have not changed.

Read More →

Oji Echo Is Live — and Your Conversations Travel Through a Third-Party Server

On September 15, 2026, Oji Echo went live. With him came a decision we owe you a straight account of: conversations with LEGH.org’s AI companions now travel through a third-party server, instead of staying on your own device. This post answers the five questions you’d ask — who, what, when, why, and where — in plain language.

What we decided

The strongest privacy design for an AI companion is one where the conversation never leaves your device or machine. That was our original plan, and it is still where we want to end up.

Read More →