AI for Genealogy: How AI Uncovered a Civil War Soldier's PTSD | Free Transcription & Research Tools for Military Genealogy

The Soldier Who Came Home Different: Using Free AI to Uncover PTSD in Civil War Records

A Union soldier survived three years of brutal Civil War combat, witnessing battles where 30% of his regiment was killed or wounded. He came home physically unscathed. But something broke inside.

His hands trembled. Thunder sent him into hiding. He woke screaming from battle nightmares. His wife said, "He's not the boy I married in 1861."

In 1884, he applied for a disability pension. Doctors documented "irritable heart," "nervous prostration," "tremor of hands." The pension bureau denied his claim. No visible wounds = no disability.

Until I used three FREE AI tools to uncover what 19th-century medicine couldn't diagnose: PTSD, documented, verifiable, heartbreaking.

In This Episode, You'll Learn:

✅ How to use Perplexity (FREE) to research military unit histories in 60 seconds with citations 

✅ How to use Gemini 3 (FREE via Google AI Studio) to transcribe 19th-century handwriting with under 2% error rate 

✅ How to use Claude (FREE tier) to analyze pension documents and spot trauma patterns doctors missed 

✅ The exact verification workflow to confirm AI insights (genealogical proof standards matter!) 

✅ Why some Civil War pensions were denied and later approved, the 1890 law that changed everything 

✅ How to recognize PTSD, depression, and anxiety in military records across ALL wars

These techniques work for ANY military ancestor:

• Civil War pension files
• WWI draft cards and service records
• WWII discharge papers and unit histories
• Korea and Vietnam era documents
• Revolutionary War applications
• Any war, any era, any handwritten military document

What Makes This Episode Different:

This isn't just Civil War research, it's about understanding the invisible wounds our military ancestors carried home and finally having the tools to honor their full stories.

You'll get copy-paste ready prompts for all three AI tools (Perplexity, Claude, Gemini 3), verification checklists, and a complete multi-tool workflow that saves hours of research time.

Every tool featured has a FREE tier. No expensive subscriptions required.

Featured Story: Private Daniel Hartley, 6th West Virginia Infantry, survived Cloyd's Mountain where his friend James Keener died in his arms. For 20 years, his community questioned whether his suffering was real because they couldn't see a wound. AI helped me find his voice, verify his testimony, and understand his trauma.

Join Our Community: Get all the AI prompts, tools guides, and research support in our FREE Facebook group "Ancestors and Algorithms: AI for Genealogy" (800+ members helping each other break down brick walls)

Resources Mentioned:

• Perplexity.ai (free research with citations)
• Claude.ai (free document analysis)
• Google AI Studio at aistudio.google.com (free transcription)
• National Archives Civil War records
• Verification workflow checklist

Connect with Ancestors and Algorithms:

📧 Email: ancestorsandai@gmail.com
🌐 Website: https://ancestorsandai.com/
📘 Facebook Group: Ancestors and Algorithms: AI for Genealogy - www.facebook.com/groups/ancestorsandalgorithms/

Golden Rule Reminder: AI is your research assistant, not your researcher.

Join our Facebook group to share your AI genealogy breakthroughs, ask questions, and connect with fellow family historians who are embracing the future of genealogy research!

New episodes every Tuesday. Subscribe so you never miss the latest AI tools and techniques for family history research.

Transcript

The pension examiner's question was simple. Did you know Private Daniel Hartley during the war? But Thomas Mitchell's answer in 1885 was anything but simple. I knew Danny before the war. I knew the man who came back. I'm not sure they were the same person. That one statement sent me down a research path that would reveal not just what happened to a Civil War soldier, but how AI helped me understand the invisible wounds that families couldn't name in the 1860s. Welcome to Ancestors and Algorithms, where family history meets artificial intelligence. I'm your host, Brian, and today we're solving a Civil War mystery that's been hiding in plain sight for 160 years. This is a story about a Union soldier from Western Virginia who survived the war physically but came home fundamentally changed. And it's a story about how three different AI tools helped me piece together what his family and doctors couldn't fully understand. The psychological cost of combat. Let's dive in. 

Here's how this started. I was researching Union soldiers from what's now West Virginia, North Virginia and I stumbled across a pension application for Daniel Hartley. Nothing unusual there. Thousands of Civil War veterans applied for pensions. But something about his file felt off. Daniel enlisted in August 1862 at age 22, joined Company F of the 6th West Virginia Infantry. Farm boy, strong as an ox according to his enlistment papers. 5 foot 9, brown hair, blue eyes. He mustered out in June 1865, three full years later. Now, you might be thinking that sounds like a success story. He survived. And you'd be right. Except for one thing. When I looked at his pension claim from 1884, nearly 20 years after the war ended, he's applying for a disability pension based on nervous exhaustion and heart disease arising from service in the war. But here's what caught my attention. 

How does a soldier serve three years with no documented injuries, but claim 20 years later that the war disabled him? And then I found that testimony from his friend Thomas Mitchell. I knew Danny before the war. I knew the man who came back. I'm not sure they were the same person. Now, I've been doing genealogy for 15 years and I've seen a lot of Civil War records. This pattern, the physical survival, the delayed pension claim, the vague language about nervous conditions. I've seen it before. This felt like what we today call PTSD, post-traumatic stress disorder. But in 1884, they didn't have that term. The question that grabbed me was, could AI tools help me understand what Daniel experienced? Could I find evidence of combat trauma in these old records that his doctors couldn't name? I had his compiled military service record, the basic facts of his service. I had his pension application from 1884. And I had something potentially powerful. Handwritten affidavits from people who knew him before and after the war. All describing this profound change, but struggling to explain it. This is where AI is your research assistant, not your researcher. The AI wasn't going to tell me what happened to Daniel, but it could help me analyze dozens of documents, research Civil War medicine, and spot patterns that might explain the unexplainable. Let me show you how this unfolded. 

First thing I needed was context. What did the 6th West Virginia Infantry actually experience during the war? Where did they fight? What kind of combat did they see? For this, I turned to perplexity. Now, perplexity is completely free to use. You get unlimited basic searches and 5 advanced searches per day on the free tier. What makes it perfect for this kind of research is that it provides citations with every answer, so I can verify what it's telling me. I asked perplexity, quote, Research the 6th West Virginia Infantry during the Civil War. What major battles did they participate in between 1862 and 1865? What was their combat experience like, end quote? Within seconds, perplexity came back with specific battles. Second Battle, Kernstown, Floyd's Mountain, Piedmont, Fishers Hill, Cedar Creek, all in Virginia's Shenandoah Valley. These weren't minor skirmishes. These were brutal, sustained combat operations. But here's what really caught my eye. Perplexity pulled up regimental histories, noting that the 6th West Virginia saw particularly heavy engagement at Cloyd's Mountain in May 1864 with high casualty rates. I asked a follow-up. Quote, what were casualty rates for the 6 West Virginia at Cloyd's Mountain? End quote. The answer? The regiment lost nearly 30% of its men, killed, wounded, or missing in a single day of fighting. Let that sink in. Daniel was there. He survived physically unscathed according to his service record. But one in three of the men around him didn't make it through that day. Now, I needed to look at his actual pension file. This is where Claude came in. I have Claude Pro, which costs $20 a month, but honestly, the free version of Claude would work great for this too. You get plenty of daily usage for genealogy research. Why Claude for this task? Because Claude excels at analyzing complex documents and spotting patterns across multiple sources. And pension files are messy. They're collections of affidavits, medical examinations, bureaucratic correspondence, all in 19th century legal language. I uploaded three documents to Claude. Daniel's pension application, the physician's examination report, and that testimony from Thomas Mitchell. Then I gave Claude this prompt. Quote, I'm researching a Civil War veteran's pension application from 1884. The veteran served 1862 to 1865 with no documented injuries, but claims disability from nervous exhaustion and heart disease. Please analyze these three documents and identify 1. All symptoms mentioned. 2. How those symptoms are described by different people. 3. Any patterns in the language used. And 4. Whether these symptoms align with what we'd now recognize as trauma response. What Claude found was fascinating. It created a table comparing how different people described Daniel's condition. The physician used terms like irritable heart, nervous prostation, and inability to perform sustained labor. Thomas Mitchell described him as startles at sudden noises, can't sleep through the night, and doesn't want to be in crowds anymore. Daniel's wife, and this got me, said, he wakes screaming from dreams of battle, and sometimes doesn't recognize where he is for several moments. Claude flagged this as matching modern PTSD criteria. Hypervigilance, sleep disturbances, flashbacks, avoidance behavior. But here's the thing. The pension examiner in 1884 didn't approve his claim initially, because without visible wounds, there was skepticism about whether the war had really caused his problems. This is where the research got really interesting. Daniel's pension file contained handwritten letters from him, from his wife, from his doctor. And some of that handwriting was challenging. 

Enter Gemini 3, which has absolutely revolutionized handwriting transcription. And here's the best part. It's completely free if you use Google AI Studio instead of the regular Gemini app. Just go to aistudio. google.com. Now, let me tell you about Gemini 3's capabilities, because this is genuinely impressive. Recent testing shows it's achieved what they call expert human-level transcription. We're talking less than 2% error rate on complex 19th century handwriting. That's better than most professional transcribers. I had a handwritten letter from Daniel to the Pension Bureau in 1885 written after his initial denial. The handwriting was shaky, inconsistent, which itself told a story. I uploaded the image to Google AI Studio and prompted, Quote, This is a handwritten letter from an 1885 Civil War pension application. Please transcribe it exactly as written, preserving original spelling and punctuation. Use bracket unclear for any words you can't decipher. This is from a veteran describing his war experiences and current health condition. End quote. Here's what Gemini transcribed. Quote, I was at Cloyd's Mountain. I seen things no man should see. My friend James Keener died in my arms. I held him while he bled and there was nothing I could do. Since then, I cannot work a full day. The sound of thunder sends me to hiding. My heart races for no reason. I wake my wife with my shouting in the night. I am not the man I was. End quote. Now, you might be thinking, wow, that's a powerful testimony. It is. But here's where verification comes in. Remember, AI is your research assistant, not your researcher. I needed to confirm that James Keener actually existed and actually died at Cloyd's Mountain. So, I went back to perplexity and searched, quote, James Keener, 6th West Virginia Infantry, Cloyd's Mountain, Casualties, May 1864, end quote. Just a quick side note. I do change names for privacy reasons. Perplexity pulled up the casualty rolls. Private James Keener, Company F, killed in action, May 9, 1864, Cloyd's Mountain. That verification changed everything. Daniel wasn't exaggerating or inventing trauma. He was accurately remembering a specific documented event. But I wanted to understand more about how Civil War medicine understood these symptoms. Back to perplexity. Quote, what did Civil War doctors know about psychological trauma and nervous conditions in soldiers? How did they treat what we'd now call PTSD? End quote. The answer was eye-opening. They had terms for it. Soldier's heart. Irritable heart. Nostalgia. Melancholia. They recognized that soldiers experienced these symptoms. But, they understood them as physical heart problems or moral weakness, not psychological wounds. The treatment? There wasn't really one. Rest? Maybe. Discharge? If it was bad enough. But, mostly soldiers were expected to just... get over it. Now, I had Claude analyze the physician's examination report from Daniel's 1885 pension claim. I asked, quote, Based on this medical report's description of irritable heart and nervous symptoms, what specific observations did the doctor make? What tests did he perform? What was his medical reasoning? End quote. Claude pulled out the key details. The doctor noted Daniel's pulse was 120 at rest, leaping to 150 with mild exertion. That's tachycardia. Abnormally fast heart rate. The doctor also noted tremor of the hands, excessive sweating even in cool weather, and patient appears constantly alert and anxious. Here's what's fascinating. Modern research on PTSD shows that these exact symptoms, elevated heart rate, hypervigilance, tremors, are psychological responses to trauma. The 1885 doctor was seeing and recording the physical manifestations of psychological trauma, even though he didn't understand the connection. But there was still a missing piece. Why was his pension initially denied if the doctor documented all of this? I went back to Claude with a different question. Quote, compare the initial pension denial letter from 1884 with the latter approval in 1890. What changed? End quote. Claude identified the key difference. The 1884 denial said, quote, no disability traceable to military service can be established. End quote. The 1890 approval came after a law change, the Disability Pension Act of 1890, which allowed pensions based on any disability, not just those provably caused by combat. Daniel didn't change. His symptoms didn't change. The law changed. And suddenly, I understood the full arc of his story. Daniel fought at Claude's Mountain, witnessed devastating casualties, including his friend's death. He came home physically intact, but psychologically shattered. For 20 years, he struggled with what we'd now call PTSD, while his community questioned whether his suffering was real because they couldn't see a wound. He tried to get help through the pension system and was initially rejected. Then, in 1890, a compassionate law change finally acknowledged that invisible wounds counted too. 

Here's what makes this story both heartbreaking and hopeful. I found one more document in Daniel's pension file. A letter from 1892, two years after his pension was approved. Jim and I transcribed it for me. It was from his wife, Sarah, writing to thank the pension bureau. She wrote, The pension means my husband can rest when he needs to. He doesn't push himself beyond what he can bear anymore. He's still not the boy I married in 1861, but he's finding some peace. Daniel lived until 1903. He was 63 years old. His death certificate listed heart failure as the cause. But I wonder if we'd diagnose it differently today. What AI helped me see was the full humanity in these documents. Without perplexity, I wouldn't have understood the specific combat experiences that traumatized him. Without Claude's analytical power, I wouldn't have spotted the pattern connecting his symptoms to modern PTSD criteria. Without Gemini's transcription ability, I wouldn't have heard Daniel's voice describing his suffering in his own words. But here's the thing, and this is crucial. The AI tools didn't solve this mystery for me. They gave me the ability to work faster, analyze more deeply, and make connections I might have missed. I still had to verify James Keener's death against the casualty rolls. I still had to research the Disability Pension Act of 1890. I still had to understand Civil War medicine in historical context. The AI tools amplified my research ability, but they didn't replace my genealogical judgment or my responsibility to verify everything. This story matters beyond just one soldier. According to pension records, tens of thousands of Civil War veterans experienced what we'd now recognize as PTSD. But because the medical understanding didn't exist, they suffered in silence, often denied help because their wounds weren't visible. Daniel's case is just one, but multiply it by thousands and you start to understand an invisible legacy of the Civil War that we're only now beginning to fully recognize. 

Okay, let's talk about what you can take from this and use in your own research. Because the techniques I used here work for lots of different genealogy challenges, not just Civil War pensions. First, the multi-tool workflow. I used three different AI tools strategically. Perplexity for historical context and citation-backed research. Completely free, unlimited basic searches. I used it to verify facts like casualty lists and understand military history. Claude for complex document analysis. Free tear works great, or $20 a month for pro if you want more intensive use. I used it to spot patterns across multiple documents and connect symptoms to modern medical understanding. Gemini 3 via Google AI Studio for handwriting transcription. Totally free and shockingly accurate. I used it to read Daniel's handwritten letters that would have taken me...

The key lesson? Match the tool to the task. Don't try to use one AI tool for everything. Second technique? Verification choreography. Every AI-generated insight needs to be confirmed against primary sources. When Gemini transcribed that James Keener died in Daniel's arm, I confirmed Keener's death through casualty rules. When Claude identified PTSD-like symptoms, I researched Civil War medical terminology to confirm doctors actually used those terms. This is that golden rule in action. AI is your research assistant, not your researcher. The AI helped me find and analyze information faster, but I made the historical connections and verified the facts. Third technique? Context loading. When I asked for complexity about the 6th West Virginia Infantry, I gave it specific date ranges and asked for combat experiences, not just general history. When I gave Claude the pension documents, I explained what I was looking for. Symptom patterns related to trauma response. The more context you provide, the better the AI can assist you. Here's your homework for this week. Find a Civil War soldier in your family tree. If you don't have one, pick any Union soldier from a state you're researching. Then try this. Use perplexity to research their regiment's combat history. Ask specific questions like, What battles did the ex-infantry participate in during year? Make note of the citations perplexity provides. Then, if you can access their pension file, either through Fold3, Ancestry, or by ordering it from the National Archives, try uploading a page to Claude and asking it to summarize the key facts and any patterns it notices. Share your discoveries in our Facebook group, Ancestors and Algorithms, AI for Genealogy. I want to hear what you find out about your Civil War ancestors. Now, I know some of you might be thinking about other wars. World War I, World War II, Korea, Vietnam. The same techniques apply. Combat trauma exists across all wars, and pension files often contain the most detailed information about a soldier's experiences. AI tools can help us analyze these records and understand the human cost of war in ways previous generations of genealogists couldn't. One more thing before we wrap up. If you're researching Civil War soldiers and you haven't explored pension files yet, start there. Military service records tell you where someone was and when. Pension files tell you who they were, how the war affected them, and how they lived afterward. They're genealogical gold mines. And if those pension files contain handwritten documents, and most do, Gemini 3 through Google AI Studio is going to save you hours of squinting at cursive. Thank you so much for listening to Ancestors and Algorithms. If you enjoyed this episode, please leave a review wherever you listen to podcasts. It helps other family historians find us. Don't forget to join our Facebook group, Ancestors and Algorithms, AI for Genealogy, where you can share your discoveries and connect with other researchers using AI. I will also be posting additional tips and techniques, along with copy-paste-ready prompts from this episode, so come over and join the group and grab those. This has been Ancestors and Algorithms. I'm your host, Brian. I'll see you next week for another journey into the past, powered by the future. Until next time, happy researching.