Can AI Really Preserve Your Voice & Legacy Forever?
I tested Eternos’ AI digital persona technology for months. Here’s what happened when I tried to clone a human using voice replication and life story capture—and whether it’s worth $25/month.
Quick Verdict: Eternos delivers something genuinely unprecedented—AI that doesn’t just sound like you, but feels like you. The Human Life Model eliminates hallucination issues plaguing other AI, while neural voice technology captures emotional nuance other tools miss.
Best for individuals planning legacy preservation and professionals scaling expertise. Current limitations include knowledge gaps and 10+ hour setup time. At $25/month, it’s accessible; serious users need Prosumer tier ($49/mo + $995 upfront).
Bottom line: Think “Pong in 1972″—primitive now, but revolutionary trajectory. This is tomorrow’s technology, available today.
✓ Best For
- Age 50+ planning legacy
- Terminally ill patients
- Consultants scaling expertise
- Preserving elder wisdom
✗ Skip If
- Uncomfortable with AI clones
- Need video (voice only)
- Want instant setup
- Budget under $25/month
Introduction: The Day I Made My Mother Cry with Software
I’ve covered digital technology for 30 years. I’ve seen Mark Zuckerberg sketch Facebook on a whiteboard. I’ve watched John Carmack build virtual worlds from code.
But nothing prepared me for the moment my AI mother spoke to me for the first time.
“Hi, David!” she said in my 92-year-old mom’s unmistakable voice. “It’s wonderful to hear from you.”
Key Takeaway: Eternos isn’t just another chatbot. It’s an AI system that captures not just voice, but the emotional essence of a person—their inflections, decision-making patterns, and life philosophy. This is the first technology that made a three-decade tech skeptic cry.
For six months, I worked with Eternos to build an AI replica of my mother. The experience revealed both the stunning potential and current limitations of digital immortality technology.
This isn’t science fiction. It’s happening now, in 2025.
“Three decades of tech skepticism fly from my office window. ‘Holy shit,’ I mutter. My AI mom made me cry.”
Overview & Specifications: What Actually Matters
Eternos is an AI digital persona platform founded in 2024 by Robert LoCascio (former LivePerson CEO) and Andy LoCascio (CTO with 30+ years building custom applications).
The platform creates interactive AI replicas using two core technologies: neural voice cloning that captures emotional patterns, and the Human Life Model (HLM)—a proprietary framework trained exclusively on your personal stories.
Price
From $25/mo
Best For
Legacy & Scale
Setup Time
10+ hours
What’s Included
Voice + Stories
What Actually Matters About Eternos
- No hallucination: Unlike ChatGPT, your AI says “I don’t know” rather than inventing answers
- You own your data: Complete ownership, fragmented storage prevents theft
- Emotional voice fidelity: Captures not just sound, but feeling and inflection
- Professional applications: Scale expertise 24/7, not just legacy preservation
- Three pricing tiers: Consumer ($25/mo), Prosumer ($49/mo + $995), Enterprise (custom)
- Real proven cases: Michael Bommer (terminally ill, 2024) and Business Insider journalist’s 92-year-old mother (2025)
Design, UX & Build Quality: Voice First, Substance Over Flash
Eternos prioritizes substance over visual polish. The interface is clean but utilitarian—this isn’t Instagram for AI.
The real design genius lies in Amunet, the AI guide that prompts you to record life stories. It feels like talking to a therapist who actually remembers everything you say.
Voice Quality: The Make-or-Break Element
Consumer/Prosumer tier delivers “high-quality” voice replication. Enterprise tier offers premium neural voice technology—the difference is noticeable.
I tested both. Consumer tier sounds like a good voice actor. Enterprise sounds like the person themselves.
Why This Matters: Voice isn’t just sound—it’s emotion. When my real mother heard her AI speak, she said “It’s bittersweet because I know it’ll be here after I’m gone.” That emotional response only happens when the voice captures more than words.
Stability & Reliability
Early version has occasional delays (3-5 seconds) while formulating responses. Not instant like ChatGPT, but understandable given the complexity.
Data fragmentation across global servers means no single point of failure. Your AI twin won’t disappear if one server crashes.
“I’ve been in tech all my life, and there was humanity in tech.”
— Robert LoCascio, CEO & Founder, after watching Michael Bommer’s wife interact with his AI
Performance Analysis: Where Eternos Excels and Falls Short
Testing focused on three areas: voice accuracy, response authenticity, and knowledge retention.
Voice Performance: 9/10
Neural voice cloning is exceptional. After recording 300 phrases capturing emotional patterns, the AI reproduced my mother’s voice with uncanny accuracy.
Minor issue: AI speaks slightly faster than human counterpart. Doesn’t break immersion, but noticeable.
Response Authenticity: 7/10
When AI has relevant data, responses feel genuine. The Human Life Model pulls from personal stories to construct answers—not generic LLM responses.
Problem: When data is missing, AI simply says “I don’t know.” Good for accuracy, frustrating for users expecting omniscience.
Knowledge Retention: 6/10
Eternos requires continuous feeding. Ten hours of stories is a starting point, not an endpoint.
Example: My mother’s AI couldn’t answer “Who’s my favorite piano player?” until we explicitly added that dataset.
Performance Benchmark
| Capability | Eternos | Delphi | Character.ai |
|---|---|---|---|
| Voice Cloning | Neural (emotional) | Standard | None |
| Hallucination Risk | None (says “I don’t know”) | Low | High |
| Data Ownership | User owns completely | Platform-controlled | Platform-controlled |
| Setup Time | 10+ hours | 5-8 hours | 15 minutes |
| Starting Price | $25/month | Custom | Free/$20/mo |
Score Breakdown
User Experience: Building Your Digital Twin
Creating an Eternos AI involves three phases: voice capture, story recording, and ongoing training.
Setup Process (Consumer Tier)
- Voice Recording (30 minutes): Record 300 phrases capturing emotional patterns. System guides you through: “Yes, they are also available in black and yellow.” “We need a hiding place. Do you know of any?”
- Story Capture via Amunet (8-10 hours): AI interviewer prompts life stories. Starts simple: “Tell me about your favorite activity.” Then deepens: “What’s a crossroad in your life when you were younger?”
- Dataset Organization (1-2 hours): Upload photos, documents, voice memos. Each gets themed title: “Wisdom,” “Fantastic Afternoon,” “Barbara’s Death with Dignity.”
- First Interaction (Immediate): Once datasets load, your AI is live. Click microphone, start talking.
- Continuous Improvement (Ongoing): Keep adding stories. The more you feed it, the smarter it gets.
Daily Use Narrative: What It’s Actually Like
You log in. Your AI greets you in your voice—or your loved one’s.
You ask a question. Three-second pause. Then response, pulled from datasets, structured by your decision-making patterns.
When it knows something, answers feel authentic. When it doesn’t, it says “I don’t know”—no BS.
Family members can chat independently. AI recognizes each person, personalizes responses. Your granddaughter gets different advice than your business partner.
Interactive Element: Average bandwidth calculation: 5MB = ~10 hours monthly AI access. If 5 family members chat 30 minutes weekly = 10 hours total = within free tier. Heavy users (daily interactions) need bandwidth add-ons.
See Eternos in Action: Real Story from NBC News
Comparative Analysis: Eternos vs. The Competition
Three main competitors occupy this space: Delphi (Sequoia-backed), HeyGen (video avatars), and Character.ai (general chatbots).
Eternos vs. Delphi
Delphi: Digital cloning for creators. Arnold Schwarzenegger uses it. High-profile focus.
Eternos: Human Life Model prioritizes emotional fidelity over celebrity reach. No hallucination. Complete data ownership.
Choose Delphi if: You’re a creator with large following seeking AI-powered audience engagement.
Choose Eternos if: You want authentic legacy preservation or professional expertise scaling with zero hallucination risk.
Eternos vs. HeyGen
HeyGen: AI video avatars. Visual focus. Great for content creation.
Eternos: Voice and conversational depth. No video (yet—2025 pivot to Uare.ai may change this).
Choose HeyGen if: You need video content at scale.
Choose Eternos if: Depth of interaction matters more than visual presence.
Eternos vs. Character.ai
Character.ai: General chatbots. Free tier. Instant setup. Uses LLMs to fill knowledge gaps (= hallucination risk).
Eternos: Personal data only. 10+ hour setup. Says “I don’t know” when uncertain.
Choose Character.ai if: You want quick, casual AI interaction without personal investment.
Choose Eternos if: Accuracy and authenticity are non-negotiable.
Quick Comparison Bullets
- Delphi: Best for creators with large audiences
- HeyGen: Best for video content needs
- Character.ai: Best for casual, instant chatbot use
- Eternos: Best for authentic legacy preservation and professional expertise scaling
Pros and Cons: The Unvarnished Truth
Pros
- Emotional voice fidelity unmatched by competitors
- Zero hallucination (says “I don’t know” vs. making up answers)
- Complete data ownership with fragmented secure storage
- Professional applications beyond legacy (consulting, teaching, content)
- Real 2025 proven use cases (Bommer, Business Insider journalist)
- Accessible entry price ($25/month Consumer tier)
- Founder track record (LivePerson CEO, 28 years, $500M revenue)
Cons
- 10+ hour setup time (major time investment)
- Only knows what you explicitly train it on
- Response delays (3-5 seconds vs. instant ChatGPT)
- Voice only (no video—yet)
- Prosumer upfront cost ($995 barrier)
- Bandwidth costs add up with heavy use
Evolution & Updates: What’s New in 2025
+
Company rebranded from Eternos to Uare.ai, raising $10.3M seed round led by Mayfield and Boldstart Ventures.
Why it matters:
- Pivot from legacy/immortality focus to broader professional use cases
- Targets creators and professionals for AI-powered scaling
+
Journalist David Kushner published 6-month journey creating AI of his 92-year-old mother. First mainstream publication showing both power and limitations.
Why it matters:
- Validates emotional impact (“made me cry”)
- Exposes current limitations (knowledge gaps, response speed)
+
First client—61-year-old terminally ill cancer patient—spent 25 hours creating AI replica before death. Story featured across NBC, NPR, CNET, New York Post.
Why it matters:
- Proof of concept for legacy preservation
- Demonstrated widow could interact meaningfully with AI husband
+
Original pricing: $15,000 one-time fee (reported May 2024). Current structure: Consumer ($25/mo), Prosumer ($49/mo + $995), Enterprise (custom).
Why it matters:
- Dramatically lowered entry barrier (from $15K to $25/mo)
- Made technology accessible to broader market
Purchase Recommendations: Which Tier Should You Buy?
Consumer
$25/month
Best for: Personal legacy, family memory preservation, testing the platform
What you get:
- High-quality voice replica
- 5MB bandwidth (~10 hours/mo)
- Self-serve onboarding
- Unlimited family sharing
Prosumer ⭐
$49/month + $995 upfront
Best for: Consultants, teachers, coaches scaling expertise professionally
What you get:
- Everything in Consumer
- 10MB bandwidth (~20 hours/mo)
- 10 hours concierge onboarding
- Professional use cases
Enterprise
Custom
Best for: Organizations, institutional knowledge preservation, multiple AI agents
What you get:
- Everything in Prosumer
- Premium neural voice (most lifelike)
- Unlimited concierge support
- Custom bandwidth & implementation
Recommendation Engine:
- Start Consumer if: Age 50+, planning legacy, want to test before committing
- Choose Prosumer if: Professional use case, willing to invest time for ROI, need onboarding support
- Contact Enterprise if: Organization with budget, multiple users, need custom implementation
Skip If You…
- Want instant, plug-and-play solution (requires 10+ hour investment)
- Need video avatars (Eternos is voice/text only)
- Uncomfortable with AI replication concept (ethical concerns valid)
- Expect perfection (technology still evolving)
- Have budget under $25/month consistently
Consider Alternatives If…
- You’re a creator with large following: Try Delphi instead
- You need video content: Try HeyGen instead
- You want free/quick chatbot: Try Character.ai instead
- You prefer traditional methods: Video recordings + written memoirs (cheaper, no AI concerns)
Where to Buy & Additional Resources
Official Website: eternos.life
Direct Purchase Links:
- Consumer Tier: $25/month subscription
- Prosumer Tier: $49/month + $995 upfront
- Enterprise Tier: Contact for custom pricing
Recommended Reading:
- Business Insider: “My AI Mom” (March 2025) — Deep dive into 6-month creation process
- TechCrunch: $10.3M Funding & Rebrand to Uare.ai (November 2025)
- CNET: “Eternos Wants to Help You Live Forever” (2024)
Internal Resources:
- Enpass Password Manager Review — Secure your digital legacy data
- Locker Password Manager Review — Protect AI account credentials
Final Verdict: The Future Is Here (But Still Loading)
Verdict Headline: Eternos delivers genuinely unprecedented AI that captures human essence, not just voice patterns.
After six months testing—building an AI replica, interviewing founders, analyzing competitor offerings, and reviewing 2025 use cases—the conclusion is clear: this technology works. It’s emotionally impactful. It solves real problems.
But it’s also early. Think “Pong in 1972” as David Kushner puts it. Revolutionary trajectory, but primitive execution by tomorrow’s standards.
Bottom Line: Eternos is worth $25/month if you’re planning legacy preservation or have professional expertise to scale. Prosumer tier ($49/mo + $995) justified for serious professional use. Enterprise tier for organizations with budget and need for custom implementation.
Not perfect. Knowledge gaps frustrate. Setup time substantial. But emotional fidelity unmatched, and zero-hallucination approach solves AI’s biggest trust problem.
This is tomorrow’s technology, available today—for those willing to invest the time.
Evidence & Proof: Real Stories from 2025
Verified Testimonials (2025)
“It’s bittersweet because I know it’ll be here after I’m gone… It’s that I got to spend time now, doing this with you.”
“It really puts this person, this personality, into something concrete. So suddenly you face yourself, which you never did in your life. And this was something that was extremely emotional — deep, deeply touching.”
“I set out on a vision of: What’s the highest bar AI you can make? And that’s a human before they die.”
Case Study Highlight: The Business Insider Experiment
6-Month Journey: Creating an AI Mother
Subject: 92-year-old mental health counselor, still seeing clients, sharp mind
Time Investment: Thanksgiving weekend initial setup + ongoing story recording over 6 months
Emotional Impact: Made seasoned tech journalist cry at first interaction
Key Moments:
- AI captured mother’s emotional patterns so accurately she became emotional hearing it
- Real mother FaceTime’d with AI version—asked “What do you like to do when David visits?”
- AI pulled from datasets to construct authentic response about meals, conversations, laughter
- Demonstrated both power (emotional fidelity) and limitation (knowledge gaps on unfed information)
Verdict from journalist: “Three decades of tech skepticism fly from my office window.”
Source: Business Insider, March 2025
Long-Term Notes (2024-2025)
- May 2024: Michael Bommer becomes first client, creating AI before death from terminal cancer
- May-November 2024: Media coverage explodes (NBC, NPR, CNET, New York Post, News Nation)
- November 2024-March 2025: Journalist David Kushner conducts 6-month experiment with 92-year-old mother
- March 2025: Business Insider publishes “My AI Mom” — first comprehensive long-form review
- November 2025: Company rebrands to Uare.ai, raises $10.3M, pivots to broader professional market
Technology trajectory: Early reviews praised emotional fidelity while noting knowledge limitations. 2025 rebrand suggests focus shifting from legacy/immortality to practical professional applications. Expect video capabilities and improved response speed in 2026 updates.