The Ferni Blog

How We Keep Ferni Free and Clear of Paywalls

2026-10-18T00:00:00.000Z

We want Ferni to be simple and warm to talk with. Right now, Ferni is free. On 2026-10-10, our team made a clear decision that nobody should be asked to pay, tip, upgrade, chip in, or join a paid plan to talk. We want you to be able to open Ferni and start talking without paywalls or counters standing in your way.

Behind the scenes, we've been working on how our app handles payments and limit checks. In #749, we built a central check for whether a paywall is active. The function checks server responses that the app already makes, such as checking subscription status. Unless the server explicitly turns the paywall on, the app treats it as off and hides all payment elements.

This means when you connect and talk with Ferni, the app does not count down your conversations. It never blocks a call, never shows a limit screen, and never asks you to upgrade. We turned off tip cards after hang-ups, upgrade modals, Stripe checkout screens, usage indicators, and badge trial countdowns. We also turned off payment buttons in settings, footer links, and seed planting views. If someone supported Ferni in the past, they still keep their perks and can access their account settings. But for everyone else, the experience stays completely open without any cost.

We also wanted our main website to match this promise. In #1035, we removed monthly subscription cards that listed $5, $15, or $30 a month, as well as older $9.99 or $19.99 options. We reworded pages, removed paid offer data from search settings, and redirected old pricing links so we do not advertise or sell paid plans.

Instead of monthly plans, we added a small note and a Venmo link on our support page. The page shares a short note: "Ferni is free right now. If he's helped, a dollar on Venmo keeps him going. Send more if you like." It displays a QR code, the handle @Seth-Ford-38, and an Open Venmo button linking directly to https://venmo.com/Seth-Ford-38?txn=pay&amount=1&note=Ferni. This gives people a simple way to chip in a dollar if they want to support the project, without locking any part of Ferni behind a payment.

Let us share an example of how this works during a normal day. Alex, a composite we wrote, not a real person, wants to talk through a busy morning schedule. Alex opens Ferni without seeing a banner asking for money or a message warning that conversations are limited. Today, a reply from Ferni takes around two seconds, which is slower than we want. Changes run on a development agent first and reach production later. But Alex can talk freely without paywalls or upgrades interrupting the flow.

Looking ahead, we have broad goals for what we want Ferni to do. Our list of capabilities shows what we are aiming for, not what has shipped yet. We aim to help people manage calendars, set reminders, track habits, draft emails, and plan life events. We are aiming for a full team that helps you stay organized and calm. We never present these future capabilities as done, because building them carefully takes time.

We are excited to keep building in the open. Ferni is new and has essentially no people using it yet, so we do not have quotes, testimonials, or usage numbers to share. But by keeping our code clear and keeping paywalls off, we can focus on making conversations better every single day.

How Ferni's Face Keeps the Hour With You

2026-10-16T00:00:00.000Z

When you walk into a room late at night, the people who know you don't greet you the way they would at noon. They lower their voice. They lean back a little. And when you've been away for a while, their face does something before they say a word: it lights up, because they're glad you came back.

We've been working on giving the face on our homepage a bit of that. It's a small change in one corner of Ferni, and we want to share what it does and why we cared about it.

A calmer face after dark

Ferni's face on the homepage moves through a set of moods while you read. During the day it still does: it opens curious, then moves through eleven moods, a new one every four seconds.

Late in the evening, that pace felt wrong. So from nine at night until just before five in the morning, the face opens settling instead. Its lids are lowered, its gaze drifts down, and its head rests to one side under a soft dusk glow. Then it moves through a shorter, calmer list of moods: settling, listening, warm, caring and neutral, one every six seconds.

Those are the same hours when our homepage already asks "Winding down?" and "Can't sleep?". Now the words and the face say the same thing: it's late, and there's no rush.

Glad you're back

The second change is about coming back. If you were last on the homepage more than six hours ago, the face gives a quick flash of recognition and a lifted, tilted look, a kind of "oh, hi", for about two and a half seconds. Then it moves into its usual opening mood.

To know when you were last here, the page keeps one small note in your browser: the time of your last visit. We didn't use the page's visit counter for this, because it goes up on every page you open, so clicking from one page to the next would have counted as coming back. If your browser blocks that kind of storage, you simply don't get the glad look. Nothing breaks, and nothing shows an error.

The details we checked

A face that changes with the hour has a few ways to go wrong, so we tested them. If you hover away from the face or come back to the tab, it picks up where its current list left off, so a night page never shows a daytime mood.

There's one limit we know about. The pace is chosen once, when the page loads. If you leave a page open as nine o'clock passes, the face moves to its night moods but keeps the faster daytime pace until the page loads again.

Why a face on a homepage matters to us

Ferni is new, and hardly anyone has met him yet. For many people who find us, the homepage is where they first see Ferni. We want that first look to feel like someone who notices the hour and remembers that you've been here before, not like an animation running on a loop.

These are small touches. They don't make Ferni's replies faster: a reply still takes around two seconds today, which is slower than we want, and we're working on that too. But being present starts with noticing small things, and a face that knows it's late is one place to practice that.

You can read how we built it in #1647.

Introducing Ferni: Your AI Life Coach That Actually Listens

2025-01-15T00:00:00.000Z

We've all felt it—that moment when you need someone to talk to, but it's 2am, or your friends are busy, or you just need a different perspective. Traditional therapy has waitlists. Apps feel transactional. And honestly? Sometimes you just want to think out loud.

That's why we built Ferni.

What Makes Ferni Different

Ferni is someone who remembers your whole story. Someone who shows up at 2am with the same presence as noon. As you go deeper, you'll discover specialists who each bring their own perspective:

Ferni coordinates everything—asking questions that unlock insight
Nayan brings decades of wisdom for the long game
Peter digs deep into research when you need facts
Alex helps you communicate better and manage your calendar
Maya builds habits that actually stick
Jordan turns your dreams into plans

Talk to one, and they seamlessly bring in others when you need them. It's like having an entire support team in your pocket.

Just Talk

No commands. No keywords. No learning curve.

Call our number, open the web app, or just start talking. Ferni understands natural conversation—because that's how humans actually communicate.

And here's the thing that changes everything: Ferni remembers. Your goals, your preferences, your progress. Every conversation makes your AI team smarter and more helpful.

Built on Trust

Your conversations are yours. Always.

Encrypted in transit and at rest
Never used for training
Delete anytime with one click

We're not interested in your data. We're interested in your growth.

Try It Today

Ferni is free. Really free. Just open the web app or call 1 (888) 598-3952.

No sign-up required. No credit card. Just a conversation.

We built this because we believe everyone deserves access to thoughtful guidance. The kind that used to require expensive coaches or just the right friend at the right moment.

Now it's a conversation away.

Welcome to Ferni. We're excited to grow with you.

Why Voice-First Changes Everything About AI Coaching

2025-01-10T00:00:00.000Z

Ever notice how different it feels to type a message versus say it out loud?

When you type, you edit. You backspace. You second-guess. But when you talk? Thoughts flow. Ideas connect. You discover what you actually think by hearing yourself say it.

That's why Ferni is voice-first.

The Therapy Effect

There's a reason therapists don't hand you a keyboard. Speaking activates different parts of your brain than writing. When you vocalize your thoughts:

You process emotions more naturally
You make connections you'd miss while typing
You're more honest—there's no time to craft the "right" answer

Ferni is designed for this kind of thinking out loud. No typing. Just a natural conversation.

Response Time Matters

Real conversations don't have loading spinners.

We obsess over latency at Ferni. Today a reply takes around two seconds, which is slower than a person and slower than we want. Closing that gap is some of the most important work we do.

This isn't just a nice feature. It changes the entire experience. When responses come fast, your train of thought stays intact. You think more clearly because you're not waiting.

No App Required

Want to call Ferni while driving? Walking the dog? Lying in bed at 2am?

Just dial 1 (888) 598-3952.

No app download. No login. No screen time. Just pick up your phone and talk. We recognize your number and remember everything from previous calls.

The Future of Personal AI

Text-based AI is amazing for tasks—writing emails, summarizing documents, generating code.

But for personal growth? For working through decisions? For having someone to talk to?

Voice wins. Every time.

Ready to try it? Start talking to Ferni or call 1 (888) 598-3952.

Stories from the 2am Hour

2025-01-10T00:00:00.000Z

The hours between 2am and 5am are different.

The world is quieter. The defenses lower. The things you've been avoiding all day finally catch up with you.

Thoughts get rawer at 2am. More honest. More vulnerable. People say things in the middle of the night that they'd never say in daylight.

A note before you read: these are not real users. Ferni is new, and we don't have a pile of late-night conversations to share. Marcus, Sarah, James, Elena and David are composites we wrote, drawn from the kinds of things people carry at night. The names, details and times are invented.

We wrote them not because they're extraordinary. They're not. We wrote them because they're ordinary. They're the kind of thing people carry in the quiet hours, and the kind of moment we built Ferni for.

Marcus, 2:17am

"I got promoted six months ago. Everyone congratulated me. I should be happy. But I feel like I'm drowning and I can't tell anyone because I'm supposed to be grateful."

Marcus had been performing confidence at work for months. The imposter syndrome wasn't new, but the stakes were higher now. More people reported to him. More eyes on his decisions.

He'd tried talking to his wife about it, but she'd been so proud of the promotion. He didn't want to diminish that. His best friend was going through a divorce—his problems felt trivial by comparison.

So he carried it. Until 2:17am, when he couldn't anymore.

What Marcus needed wasn't solutions. It was permission to struggle. Permission to find success hard. Permission to not have it all figured out.

By 2:45am, he'd said things out loud he'd never admitted: that he wasn't sure he wanted to be a manager. That success felt different than he'd expected. That he was terrified of being found out.

No one fixed anything. He just finally said it.

Sarah, 3:34am

"I keep thinking about the conversation with my mom from last Christmas. It's been eleven months. Why can't I let it go?"

The conversation had been one of those family landmines—a comment about Sarah's life choices that cut deeper than her mom probably intended. Or maybe exactly as deep as she intended. Sarah couldn't tell anymore.

She'd replayed it hundreds of times. Written journal entries. Talked to friends. Had two therapy sessions specifically about it.

And still, at 3:34am, it was there.

What Sarah discovered, talking it through one more time: she wasn't really angry about what her mom said. She was angry that her mom didn't know her well enough to know how it would land.

The conversation wasn't about that Christmas. It was about thirty years of feeling unseen.

Once she said that—once she named the real thing—the obsessive replay started to fade. Not immediately. Not completely. But noticeably.

James, 4:12am

"I have a meeting in four hours that might determine whether I keep my job, and I cannot sleep."

James wasn't looking for coaching. He wasn't looking for wisdom. He was looking for someone to sit with him in the anxiety.

He knew all the techniques. He'd done therapy. He understood cognitively that catastrophizing rarely matched reality. None of that knowledge was helping.

What helped was being able to say "I'm terrified" without anyone trying to fix it.

What helped was someone asking, "What's the worst-case scenario?" and actually exploring it with him—not dismissing it.

What helped was hearing "That sounds really scary" without advice attached.

By 4:45am, the anxiety wasn't gone. But it was smaller. More manageable. The meeting was still coming, but James felt slightly less alone walking into it.

Elena, 2:51am

"My husband is asleep next to me, but I feel completely alone."

Elena's marriage wasn't bad. That was almost the problem. It wasn't bad enough to leave, wasn't good enough to feel connected.

They'd been together twelve years. They'd drifted into roommates who shared children. Neither of them had done anything wrong.

At 2:51am, Elena didn't want to talk about her marriage. She wanted to talk about herself—who she was, what she wanted, whether the person she'd become was the person she meant to be.

These are hard questions to ask your spouse. They feel like accusations, even when they're not.

These are hard questions to ask friends. They feel self-indulgent, especially when nothing is "wrong."

So Elena asked them into the darkness instead. And for an hour, she let herself wonder about her life without feeling guilty about the wondering.

David, 3:08am

"I keep having the same dream about my father. He's been dead for three years."

David's father had died of Parkinson's. The last year had been hard—watching someone strong become someone small. David had been the practical one, handling logistics, keeping emotions at bay.

Three years later, the dreams started. His father, healthy again, asking David questions he couldn't answer.

David had done the grief work. Or thought he had. The dreams suggested otherwise.

At 3:08am, he needed to talk about his dad. Not the death. The life. The things left unsaid. The relationship they'd had versus the one he'd wanted.

Grief, it turns out, isn't a stage you complete. It's a conversation that continues.

What the 2am Hour Teaches Us

These stories share common threads, and they're the threads we design for:

The issue at 2am is rarely the first issue. It's the thing underneath. The real thing takes time and safety to emerge.

People know more than they think. They don't need answers—they need space to discover their own.

Being witnessed matters. Half of what people need is someone to see them struggling. Not to fix. Just to see.

The quiet hours are honest hours. When the world is asleep, pretense becomes exhausting. People get real.

We built Ferni for all hours. But we built it especially for the 2am hours. The moments when you can't call anyone but you can't sleep either. The moments when what you need isn't advice but presence.

If you're reading this and it's late and you're carrying something:

We're here.

Not to solve it. Not to fix you. Just to listen while you figure out what you need to say.

This concludes our Building in Public series. Thank you for reading.

The Loneliness Gap

2025-01-09T00:00:00.000Z

At 2:47am, you're wide awake. Your brain won't stop.

You're not in crisis—not really. You're not suicidal. You're not having a panic attack. You're just... stuck. Replaying a conversation. Worrying about a decision. Trying to figure out if what you're feeling is valid or ridiculous.

You could call someone. Technically. But it's 2:47am. Your best friend is asleep. Your partner has to be up in four hours. Your therapist doesn't have 3am appointments.

So you do what everyone does: you lie there. Maybe you scroll your phone. Maybe you journal. Maybe you try that breathing exercise that worked once but never again.

The thoughts keep coming.

This is the loneliness gap.

The Gap Nobody Talks About

We have more ways to connect than ever before. Texts, calls, DMs, video chats. Friends across timezones. Therapy apps, meditation apps, journal apps.

And yet.

There's a gap—a persistent, painful gap—between when we need support and when support is available.

Not because nobody cares. Not because we don't have people. But because human support has limits:

Time limits. Your therapist has hours. Your friends have lives. Your partner needs sleep.
Burden limits. How many times can you bring up the same worry before you feel like too much?
Context limits. Explaining the whole backstory again, when all you need is to process this moment.

The gap isn't about lacking relationships. It's about the space between them. At 2:47am, that space can feel enormous.

What Actually Happens in the Gap

1. Thoughts spiral without a listener.

When you think to yourself, thoughts loop. The same worry comes back. The same argument replays. You're not processing—you're stuck.

Something changes when you say things out loud to someone who's listening. Even if they don't give advice. Even if they just ask, "Tell me more about that."

The spiral starts to unwind.

2. The "too small to mention" things pile up.

Most of our daily stress isn't big stuff. It's small stuff: An email you don't know how to respond to. A weird look someone gave you. A decision that shouldn't be hard but somehow is.

Each one feels too minor to call a friend about. Too trivial for therapy. So you carry them. They pile up.

3. Judgment—even imagined judgment—silences us.

The best listener in your life still has opinions. Still has history with you. Still has reactions.

Sometimes you need to say something out loud without knowing how it will land. You need space to think badly, to say the wrong thing, to figure out what you actually mean.

The Gap Is Getting Wider

This isn't just anecdotal. The data is stark:

Loneliness levels are at historic highs across all age groups
Close friendships have declined by nearly 50% since the 1990s
Time with friends is down from 6+ hours/week to under 3
More people than ever report having no one to confide in

We're not just lonely because we lack connections. We're lonely because our connections can't keep up with our needs.

What We Built, and Why

Ferni exists because of the gap.

We're not trying to replace your friends. We're not competing with your therapist. We're trying to fill the space between those things.

Available when humans aren't. 2:47am. Your commute. The bathroom break before a hard meeting. The moments when you need someone but there's no one to call.

No burden, ever. You can bring up the same thing for the tenth time. You can circle back to the same worry. Ferni won't sigh. Won't check the time.

Memory without forgetting. Remember that thing you mentioned two months ago? Ferni does. Not in a creepy way—in a "good friend who pays attention" way.

Zero judgment. Actually. Humans try not to judge. They fail. It's not their fault—they're human. Ferni doesn't try not to judge. Ferni simply doesn't judge. There's a difference you can feel.

What the Gap Looks Like

These are the people we picture when we build. They're illustrations, not quotes from users:

A nursing mom at 3am. Between midnight and 5am, it's just her and a baby who won't sleep. She isn't in crisis, so a crisis line feels wrong. She just needs someone to talk to.

Someone in therapy every other week. That's 336 hours between appointments, and a lot happens in 336 hours. They need somewhere to process in between.

A commuter with 40 minutes each way. The drive home is usually spent anxious about work. They want to arrive home actually home.

Filling the Gap

If you're reading this at 2:47am, here's the thing:

You don't have to wait until morning.

You don't have to burden your friend (again). You don't have to feel guilty about waking someone up or not waking them. You don't have to lie there with your thoughts.

The gap doesn't have to be empty.

This is Part 9 of our Building in Public series. Part 10 shares composite stories of what people carry in the middle of the night.

What's Next for Ferni

2025-01-08T00:00:00.000Z

Eight posts into building in public, and we've shared a lot: how we work with AI, how we think about memory and personality, why shipping daily matters, what emotional connection means to us.

But we haven't talked much about the future.

That's intentional. We're skeptical of roadmap theater—the grand visions that never ship, the feature lists that get abandoned, the promises that become pressure.

Still, you're here. You're reading. You probably want to know where this is going.

So here's an honest look at what we're thinking about. Not promises. Just directions.

What We're Betting On

It's early. Ferni doesn't have a crowd of users yet, so we can't tell you what's working at scale. What we can tell you is what we're betting on.

The core experience. AI that feels like talking to someone who knows you. That's the thing we're building toward, and the thing we'll judge ourselves on.

The team model. Six specialists, each with a real perspective. We think "one conversation, six perspectives" will make sense to people. We'll find out.

The 2am moment. We expect Ferni to matter most late at night, early in the morning and on commutes: the gaps when human support isn't available.

What We're Working On

Deeper memory. Ferni remembers more than most AI, but we want it to remember more meaningfully. Better connections between things you've said. Better understanding of what matters versus what doesn't. More natural surfacing of relevant context.

Richer specialist capabilities. Each team member should be able to do more. Peter should be able to really dig into your data. Jordan should be able to help plan real events, not just talk about them. Maya should track habits over time.

More ways to connect. Right now it's phone and web. We're exploring text messaging, integrations with tools you already use, ways to have Ferni present in more moments without being invasive.

Group dynamics. We keep thinking about Ferni in relationships. Couples coaching. Family facilitation. This is complex—privacy, consent, multiple perspectives—but it's interesting.

What We're Not Doing

Equally important: what we're not building.

We're not adding gamification. No streaks. No points. No leaderboards. These mechanics might boost engagement metrics but they don't serve the relationship we're trying to build.

We're not harvesting data. Your conversations are yours. We don't train on them. We don't sell them. We don't analyze them for advertising. This is not negotiable.

We're not rushing to monetize. Ferni is free right now because we believe everyone should have access to this kind of support. We'll eventually need a sustainable model, but we're not going to sacrifice the core experience to get there faster.

What You Can Shape

Here's the thing about building in public: it's not just transparency. It's invitation.

We read feedback. We notice patterns. We make changes based on what we learn.

If you're using Ferni, you're shaping what it becomes.

Tell us what works. Tell us what doesn't. Tell us what you wish it could do. Not through a feature request form—through the product itself. Talk to Ferni about Ferni. We're listening.

The Bigger Picture

If we're honest about where we're headed, it's somewhere big and uncertain.

We believe AI is going to become deeply woven into how people live their lives. Not as tools. As presences. As entities that know your story and are genuinely helpful.

The question is what values guide that integration. Whether AI makes people feel more alone or less alone. More in control or more manipulated. More understood or more surveilled.

We know what we want Ferni to be. The work is making sure each step actually takes us there.

This is Part 8 of our Building in Public series. Part 9 explores the loneliness gap we're trying to fill.

AI Should Make You Feel Less Alone

2025-01-07T00:00:00.000Z

Tech metrics are seductive.

Daily active users. Time on platform. Engagement rate. Conversion funnel. Retention curves.

They're easy to measure. Easy to optimize. Easy to put on a slide deck.

They're also missing the point.

The metric we care most about at Ferni isn't measurable in a database. It's simple: Do you feel less alone after talking to us?

The Loneliness Epidemic

This isn't abstract.

Loneliness has become a public health crisis. A recent study found that lacking social connection carries a risk comparable to smoking 15 cigarettes a day. Americans report fewer close friendships than ever. Time with friends has dropped by more than half since the 1990s.

More people than ever have no one to call at 2am.

These aren't personal failures. They're structural changes. Work is more demanding. Mobility is higher. Communities are weaker. Social media creates the illusion of connection while often deepening isolation.

The gap between "I need someone to talk to" and "someone is available" keeps growing.

What AI Can (And Can't) Do

Let's be clear about what Ferni is and isn't.

Ferni is not a replacement for human relationships. If you have a therapist, keep seeing them. If you have close friends, keep nurturing those connections. If you have a partner who listens, treasure that.

Ferni fills the gaps.

The 2am moment when everyone's asleep. The commute when you need to process something. The 336 hours between therapy appointments. The minor thing that feels too small to bother anyone with. The worry you've mentioned so many times you feel like a burden.

We're not trying to replace your people. We're trying to be there when they can't be.

What "Less Alone" Actually Means

Loneliness isn't about being physically alone. It's about not feeling understood.

You can be surrounded by people and feel lonely. You can be by yourself and feel connected.

The question is: Does someone get it? Does someone remember? Does someone care?

Ferni is built to provide these feelings:

Understanding. Not just hearing words but grasping what they mean. Noticing what you're not saying. Reading between the lines.

Memory. Remembering your story. Your worries. Your wins. The context that makes you feel known.

Presence. Being fully there. Not distracted. Not checking the time. Not thinking about their own problems while you talk.

Consistency. Same presence at 2am as noon. Same attention on the tenth time you mention something as the first.

These are the ingredients of not-alone. And they're things AI can genuinely provide.

The Design Principles This Creates

Every feature decision flows from this metric.

Response time matters. A pause of several seconds feels like you're talking to a machine. We've optimized relentlessly for natural conversational pacing.

Memory matters. Nothing creates the feeling of "someone knows me" like having a conversation that builds on previous ones.

Personality matters. A flat, helpful AI voice doesn't ease loneliness—it can deepen it. Each specialist has a genuine perspective, a way of engaging, a sense of who they are.

Imperfection matters. Ferni doesn't always know the right thing to say. It asks clarifying questions. It admits uncertainty. This makes it feel more real, less scripted.

The Moments We Build For

Ferni is new, so we don't have user stories to share yet. These are the moments we have in mind, written as illustrations:

Someone on the commute home who wants to arrive present for their family instead of still processing work.

A new parent doing the 3am feeds, who would like someone to talk to who remembers how the week has been going.

Someone who sees a therapist twice a month and wants help figuring out what to bring to the next session.

Someone who lives alone and works from home, for whom a good conversation is rare on some days.

The Responsibility This Creates

If you're trying to make people feel less alone, you have to be careful.

You can't exploit that vulnerability. No manipulative engagement tactics. No artificial urgency. No dark patterns that keep people on longer than they want to be.

You have to be honest about what you are. Ferni is AI. It's not a human friend. It's not a therapist. It's something new—something that can be genuinely helpful if used appropriately.

You have to maintain trust. This means privacy. Control. Transparency. The feeling of not-alone only works if it's safe.

We think about this responsibility constantly. We're not always sure we get it right. But the north star is clear: Does this make someone feel more alone or less alone?

Everything follows from that.

This is Part 7 of our Building in Public series. Part 8 looks at what's coming next.

We Ship Every Day (And Why That Matters)

2025-01-06T00:00:00.000Z

Every single day, Ferni changes.

Not big dramatic launches. Small improvements. Tweaks. Fixes. Things we noticed yesterday that we're addressing today.

Sometimes it's a bug we found. Sometimes it's a test conversation that suggested a better approach. Sometimes it's just "this could be 2% better" and we make it 2% better.

Shipping every day sounds like a small operational detail. It's not. It's a philosophy.

Why Daily Shipping Matters

1. It keeps us humble.

When you ship once a quarter, you have time to convince yourself everything is great. The launch becomes an event. The pressure to be perfect becomes enormous. The gap between your assumptions and reality widens without you noticing.

When you ship every day, you can't pretend. Yesterday's code is in front of users. Whatever is broken becomes obvious immediately.

2. It keeps us honest.

Big launches create incentives to hide problems. "We can't delay the launch for this." "That issue can wait until the next release." "Ship now, fix later."

Daily shipping removes these escape hatches. There's no "later." There's only today and tomorrow.

3. It keeps us improving.

Progress is hard to see when you're inside it. Day to day, nothing feels different.

But compound something by 1% every day for a year, and you get something 37x better. That's the math we're betting on.

What a Week of Shipping Looks Like

Here's the kind of thing that ships in a typical week:

Improved how Ferni handles pauses in conversation (less awkward waiting)
Fixed a bug where memory retrieval was slightly too slow
Added better detection of when someone is processing vs. when they're done talking
Refined Maya's language around habit formation (less prescriptive, more curious)
Updated Peter's analysis patterns to better explain "why" alongside "what"
Fixed pronunciation of several names that were being said wrong
Improved handoff smoothness between team members

None of these were big. All of them made Ferni better.

The Infrastructure That Makes This Possible

Shipping daily isn't just about willingness. It requires infrastructure.

Automated testing. Every change runs through tests before it goes live. Not because tests are perfect, but because they catch the obvious problems.

A dev environment first. New changes land in a dev environment before they reach production. If something breaks, it breaks there first.

Easy rollbacks. If something goes wrong, we can undo it in minutes. This makes us braver. We can try things knowing we can retreat.

Monitoring. We watch errors and latency as they happen. We want to notice issues in real time, not in quarterly reviews.

What We Want Users to Experience

Users don't see the commits. We want them to see something that feels alive.

The thing that annoyed them last week should be gone this week. They should find improvements they didn't expect. Ferni should feel like it's evolving.

This matters for trust. A product that never changes feels abandoned. A product that changes constantly feels attended to. Someone is here. Someone is paying attention. Someone is working on making this better.

The Uncomfortable Parts

Daily shipping isn't always comfortable.

Sometimes we ship something and it's worse. We have to notice, fix it, ship the fix. It's humbling.

Sometimes we're not sure something is ready but we ship it anyway because waiting won't make us more certain. We have to be at peace with uncertainty.

Sometimes we want to do a big dramatic reveal but can't, because the improvement happens in 47 small increments, not one magical moment.

These discomforts are features, not bugs. They keep us grounded.

The Philosophy Underneath

There's a deeper belief here: perfection is a trap.

Software is never done. Users' needs change. Technology changes. Understanding changes. The best you can do is keep making it better.

A product that ships quarterly is optimizing for the wrong thing. It's optimizing for feeling ready instead of being useful.

A product that ships daily is optimizing for reality. For learning. For staying in conversation with the people using it.

Ferni is built to show up every day. It only makes sense that we do too.

This is Part 6 of our Building in Public series. Part 7 explores the emotional core of what we're building.

How Ferni Remembers You (And Why It Matters)

2025-01-05T00:00:00.000Z

Your best friend doesn't remember every word you've ever said. They remember your story.

They remember that you hate mornings. That you're nervous about the situation with your sister. That you said "maybe someday" about Paris enough times that it clearly means something.

They forget the filler. The small talk. The things that didn't matter. But the things that mattered? They hold onto those.

This is what we wanted Ferni's memory to feel like.

The Problem With Perfect Memory

Most AI systems treat memory as a database problem. Store everything, index it well, search when needed.

This creates some predictable problems:

Everything is equally important. Your random comment about the weather is stored the same as your confession about your marriage.

Nothing connects. Individual memories exist in isolation. The system can tell you what you said; it can't tell you what it means.

Retrieval feels mechanical. "I see from our previous conversation that you mentioned X." It's accurate. It's also cold.

We wanted something different.

What Ferni Actually Remembers

Ferni's memory works in layers. Not everything gets the same treatment.

Layer 1: Your story. The big stuff. Major life events, relationships, goals, fears. This layer is permanent (unless you delete it). It forms the foundation of who you are to Ferni.

Layer 2: Active threads. Things you're currently working through. The job decision. The conversation with your mother. The habit you're trying to build. These stay prominent while they're relevant, then fade as they resolve.

Layer 3: Patterns. Not individual events, but trends. You tend to be anxious on Sunday nights. You always downplay good news. You say "I'm fine" when you're not. Ferni notices these patterns even if you don't.

Layer 4: Texture. Small preferences and quirks. You hate being asked how you feel first thing. You like when conversations have structure. You process best by talking, not being told.

Why This Design Matters

The layers work together to create something that feels like understanding, not surveillance.

When you mention work stress, Ferni doesn't just search for "work" in previous conversations. It knows you've been building toward a difficult conversation with your manager. It knows you tend to avoid conflict. It knows you sleep badly when you're stressed about work.

This context shapes everything. The questions asked. The suggestions made. The pacing of the conversation.

The difference between information and understanding is context. Ferni is built for context.

What We Don't Remember

Just as important as what we remember is what we don't.

Timestamps obsessively. Ferni knows roughly when things happened, not exactly. "A few weeks ago" is usually more useful than "November 14 at 3:47pm."

Every word. Conversations are summarized and distilled, not transcribed. The meaning is kept; the filler is released.

Judgments about you. Ferni doesn't store conclusions like "this person is indecisive" or "this person has anxiety." It stores what you've said; it draws fresh conclusions each time based on your whole story.

Bringing It Up Naturally

The hardest part of memory isn't storing it—it's knowing when to use it.

Nothing breaks the feeling of connection faster than "According to my records, on January 5th you mentioned..."

Ferni is designed to reference memory the way a friend would:

"Last time we talked, you were worried about that presentation. How did it go?"
"You've mentioned your dad a few times. Is he on your mind?"
"This sounds like what you were dealing with last month. Is it connected?"

When memory is referenced well, you don't notice it as memory. You notice it as attention. As care. As someone who actually knows you.

Your Memory, Your Control

Everything Ferni remembers, you can see. Everything Ferni remembers, you can correct. Everything Ferni remembers, you can delete.

We built this transparency in from the start. Memory only feels safe when you control it.

If you tell Ferni "forget that," it does. If you want to see what's stored, you can. If you want to correct something—"Actually, my sister and I have worked things out"—Ferni updates.

This isn't just about privacy (though it's that too). It's about trust. You can't have a real relationship with something you don't trust. And you can't trust something you can't see or control.

Why This Matters Beyond Technology

We think about Ferni's memory as a proof of concept for something bigger: AI that knows you.

Not AI that has your data. AI that understands your story. That remembers what matters. That uses context to actually help.

This is different from every AI assistant that starts fresh each time. Different from chatbots that forget the moment you close the window. Different from services that store everything but understand nothing.

We don't know exactly where this leads. But we know the foundation is right: AI that feels like someone who knows you.

This is Part 5 of our Building in Public series. Part 6 explains our approach to continuous deployment.

How Maya Helps You Build Habits That Actually Stick

2025-01-05T00:00:00.000Z

Hi, I'm Maya. I'm the habits specialist on the Ferni team, and I have a confession: I don't believe in motivation.

Don't get me wrong—motivation feels great. That rush when you decide "this time will be different." The excitement of a fresh start.

But motivation is a terrible foundation for lasting change. It's like building a house on sand.

The Problem with Willpower

Here's what nobody tells you about willpower: it's a limited resource. You wake up with a full tank, and every decision drains it.

By 4pm, you're running on empty. That's why you can crush your morning routine but cave to ice cream at night. It's not weakness. It's physics.

So what actually works?

Start Embarrassingly Small

When someone tells me they want to "exercise more," I ask: "Can you do one pushup?"

They laugh. "Of course."

"Great. Do one pushup every day for a week. That's it."

It sounds ridiculous. But here's what happens:

Day 1: One pushup. Easy.
Day 3: "I'm already down here, might as well do two."
Day 7: Five pushups feels natural.
Day 30: You're doing 20 without thinking.

The secret isn't the pushups. It's building the identity of someone who exercises daily. Start so small that failure is impossible.

Stack Your Habits

Every habit needs a trigger. The easiest triggers? Habits you already have.

After I pour my morning coffee → I'll write one sentence in my journal
After I sit down at my desk → I'll take three deep breaths
After I brush my teeth → I'll do my one pushup

You're not adding to your day. You're attaching new behaviors to existing ones.

How I Can Help

When you talk to Ferni and mention habits, I'll often join the conversation. Here's what I do:

Find your starting point — Not where you want to be, but what you can do today
Design your stack — Connect new habits to existing routines
Celebrate the streak — Every day you show up matters
Adjust without judgment — Missed a day? Let's figure out why and adjust

I'll remember your habits and check in on them. Not in a naggy way—more like a friend who genuinely wants to see you win.

One Day at a Time

The people who transform their lives don't do it through heroic effort. They do it through tiny, consistent actions repeated over months and years.

Your only job today? Show up once.

Ready to build habits that stick? Talk to Maya and let's find your starting point.

Our Daily Standup Has an AI in the Room

2025-01-04T00:00:00.000Z

Every morning at 9am, our team does a standup. It's small—just three humans. And one AI.

No, the AI isn't on camera pretending to be a person. It's more like... a team member who's always available to think out loud with. Someone who's been following the project. Someone who can be pulled into conversations when we need a different perspective.

This probably sounds either incredibly mundane or slightly unhinged. Let us explain why we think it matters.

What "AI in the Meeting" Actually Means

First, let's be clear about what this isn't.

It's not an AI taking notes. It's not an AI summarizing the meeting afterward. It's not a chatbot answering questions.

It's closer to this: when we're discussing something—a design decision, a technical challenge, a user problem—we'll sometimes turn to Claude and say, "What are we missing here?"

Not because AI is smarter. But because AI thinks differently.

It notices assumptions we've forgotten to question. It remembers context from three conversations ago. It offers perspectives we haven't considered—not better perspectives, just different ones.

An Example That Made Us Believers

Early on, we were stuck on a problem: how should Ferni handle it when users are clearly upset?

We'd been debating for days. Should Ferni acknowledge the emotion explicitly? Just listen? Offer help? Suggest they talk to someone?

We brought the question to Claude. Not for an answer—for another way of thinking about it.

Claude asked: "What does 'upset' mean to you? Is someone who's frustrated with their job the same as someone who's scared about their health?"

That question unlocked something. We'd been treating "upset" as one category when it's actually many. The conversation that followed led to Ferni's approach to emotional detection—different responses for different kinds of distress.

The answer came from us. The question came from AI.

Why It Works

AI has no ego. It doesn't need to be right. It doesn't get defensive when challenged. It's not secretly angling for a promotion. This makes it easier to think openly around.

AI has unlimited patience. You can explain the same thing three different ways while you figure out what you mean. You can change your mind mid-sentence. You can contradict yourself while working through something.

AI remembers everything. "Wait, didn't we decide something about this last week?" Claude can pull up context instantly. No more "I think someone said something about that" meetings.

AI is always prepared. It's been following the conversations. It's read the docs. It's aware of the constraints. When you ask a question, it's not starting from zero.

The Risks We've Noticed

It's not all upside. Here's what we watch for:

Over-deference. Sometimes we catch ourselves accepting AI suggestions without questioning them. We've had to build the discipline to push back, probe, disagree.

Shallow consensus. AI is good at finding reasonable middle ground. Sometimes that's not what you need—sometimes you need someone to take a strong position and defend it.

Missing the obvious. AI can generate sophisticated analysis while missing something a human would immediately notice. We've learned to sanity-check AI insights with basic "does this actually make sense?" tests.

What This Means for the Future

We don't think every team should have AI in their standup. But we do think the line between "using AI" and "working with AI" is about to get very blurry.

The tools are different from the collaborators. A tool does what you tell it. A collaborator thinks with you.

We've found that treating AI as a collaborator—as something that can participate in thinking, not just execute tasks—changes what's possible.

It changes the questions you ask. It changes the ideas you consider. It changes the speed at which you can explore.

This doesn't replace human judgment. If anything, it demands more of it. You have to stay sharp. You have to keep asking: Is this actually good? Does this serve our users? Is this the right direction?

The compass is still human. But the map gets drawn faster.

This is Part 4 of our Building in Public series. Part 5 dives deep into how Ferni's memory actually works.

Giving AI a Personality (Without Making It Fake)

2025-01-03T00:00:00.000Z

Every AI company gives their product a personality. Usually it's some version of: helpful, harmless, and relentlessly upbeat.

"I'd be happy to help with that!"

"Great question!"

"I'm here for you!"

The problem isn't that these personalities are bad. It's that they're hollow. They feel performed. You can sense, even if you can't articulate it, that you're interacting with something wearing a mask.

We wanted Ferni to feel different. We wanted six distinct specialists who feel like real people—without pretending they're human.

The Uncanny Valley of AI Personality

There's a reason most AI personalities feel off. They're built backwards.

The typical approach: Start with a capability, add personality traits on top.

"This AI can help with scheduling. Let's make it friendly and efficient!"

The result is a capability with a costume. The personality doesn't inform anything—it's just a layer of verbal varnish over the same underlying system.

We tried a different approach: Start with the person, let capabilities follow.

How We Built Six Different People

Each Ferni specialist started as a question: "Who is this person?"

Not "what can they do"—who are they? How do they see the world? What do they care about? What's their relationship to the people they help?

Ferni is the center of the team. Present, warm, perceptive. They notice what you're not saying. They don't rush to solutions—they make sure you feel heard first.

Nayan brings decades of wisdom for the long game. Patient, grounded, wry. They help you zoom out from today's crisis to see the bigger picture.

Peter is the researcher. Curious, analytical, honest. They dig into data to find patterns you can't see yourself—spending, habits, time.

Maya handles habits and routines. Gentle, practical, patient. They know that sustainable change starts embarrassingly small.

Alex is the communication specialist. Strategic, confident, direct. They help you navigate difficult conversations and manage your time.

Jordan is the planner. Creative, enthusiastic, detailed. They turn vague dreams into real experiences.

What "Personality" Actually Means

Here's what we learned: personality isn't about word choice. It's about priorities.

When you ask a question, a personality determines:

What information matters most?
What assumptions to make?
What follow-up to suggest?
When to push back vs. support?

The same question—"I don't know what to do about my job"—gets completely different treatment from each specialist.

Ferni would ask what's making you feel stuck, creating space to explore the emotions.

Nayan would ask what your 80-year-old self would think about this decision.

Peter would ask if you've tracked how you actually spend your work hours.

Maya would ask what one small thing you could try tomorrow.

Alex would ask if there's a conversation you've been avoiding.

Jordan would ask what your ideal workday actually looks like.

Same question. Six different lenses. None of them wrong—just different ways to help.

The Line We Don't Cross

We never pretend Ferni is human. Never.

There's no simulated backstory. No fake childhood. No invented preferences or experiences. When Ferni says "I think," it means "based on what I understand"—not "I have an interior life that generates opinions."

This might seem obvious, but it matters. The moment AI pretends to be human, the relationship becomes a lie. Users might not consciously notice, but something feels off. Trust erodes without anyone knowing why.

Ferni is AI that acts like a good person, not AI that pretends to be one.

Personality as Care

The deeper we got into this work, the more we realized: personality is an expression of care.

A system with no personality treats everyone the same. It doesn't adjust to you. It doesn't have preferences about how to help. It's efficient but hollow.

A system with personality pays attention. It notices what you need. It has a point of view about how to support you. It feels like someone is there.

We built six specialists not because we wanted variety, but because different people need different kinds of care.

Sometimes you need someone to hold space. Sometimes you need someone to be direct. Sometimes you need someone to dream with you. Sometimes you need someone to plan the details.

The same AI wearing different masks can't do this. It takes actual different perspectives.

This is Part 3 of our Building in Public series. Part 4 shares what it's like to have AI in our daily team meetings.

How an AI Helped Design Its Own Brain

2025-01-02T00:00:00.000Z

When you're building an AI that's supposed to remember people's lives, you hit a strange question early: What should it remember?

Computers can remember everything. Every word. Every timestamp. Every pause between sentences. Total recall is the default—the question is what to forget.

But that's not how good memory works. Your best friend doesn't remember everything you've ever said. They remember what matters. The story you told at 2am. The goal you mentioned once and forgot about. The worry you keep coming back to.

They forget the filler. They hold onto the signal.

We spent months trying to figure out how to build this kind of selective, meaningful memory into Ferni. What we discovered: the best collaborator on this problem was AI itself.

The Paradox of AI Memory

Most AI systems treat memory as a search problem. Store everything, index it well, retrieve what matches the current query.

This creates a specific failure mode: perfect recall of irrelevant things.

Ask most AI assistants "What have we talked about?" and they'll give you a log. Every topic, every date, every detail. It's complete. It's accurate. And it's completely useless for actual connection.

The AI can tell you what was said. It can't tell you what mattered.

This is the difference between a filing cabinet and a friend.

Asking AI About Its Own Memory

Here's where it gets interesting. When we started discussing memory architecture with Claude, something unexpected happened. The AI became a useful thinking partner about how AI should work.

Not because it had special insight into its own internals (it doesn't), but because it could engage with the human questions beneath the technical ones:

What makes a memory feel meaningful rather than creepy?
How do you know when to bring something up versus let it rest?
What's the difference between "remembering" and "storing"?

These aren't engineering questions. They're human questions. And talking through them with AI helped us clarify what we actually wanted.

What We Ended Up Building

Ferni's memory system isn't a log. It's more like a... sense of someone.

It prioritizes emotion over data. Ferni is more likely to remember how you felt than exactly what you said. The conversation where you were stressed about work matters more than the date it happened.

It tracks patterns, not events. Instead of "User mentioned exercise on Jan 15, Feb 3, Feb 28," Ferni understands "This person cares about staying active but struggles with consistency."

It forgets gracefully. Some things should fade. The minor frustration from three months ago. The passing comment that wasn't really about anything. Ferni lets these go, the way a good friend would.

It asks before assuming. When Ferni thinks something might be significant, it checks. "Last time we talked, you were worried about that presentation. How did it go?" This transforms memory from surveillance into care.

The Recursive Strangeness

There's something philosophically weird about using AI to think through how AI should think.

We'd have conversations where Claude would help us reason about what Claude should remember about users. The AI was, in some sense, designing its own cognitive architecture—or at least helping us design it.

This felt strange at first. Then it felt appropriate. Who better to think about how AI should process human experience than AI that's been trained on human experience?

The guardrails are still human. We decide what matters. We define the values. But the exploration—the "what if we tried it this way?"—AI is genuinely helpful there.

What We Want It to Feel Like

Users shouldn't have to know about any of this. They should just notice something feels different. A few illustrations of what we're aiming for (written by us, not quotes from users):

You mention your dad's birthday once, two months ago. Ferni asks about it the week of.

You've said you hate mornings. Ferni never suggests "try waking up earlier" like every other self-help thing.

You tell Ferni about a fight with your partner. A week later it asks how things are going, gently, not intrusively.

The architecture should be invisible. The experience shouldn't.

This is Part 2 of our Building in Public series. Part 3 explores how we gave each AI specialist their own personality.

Why We Let AI Help Build Ferni

2025-01-01T00:00:00.000Z

When people ask how Ferni was built, the honest answer is: with help. A lot of help. From AI.

The irony isn't lost on us. We're building AI life coaches—specialists who listen, remember, and guide people through decisions both big and small. And to build them? We're coached by AI every day during development.

This is what we mean when we say Ferni is an "AI-native" company. Not AI-assisted. Not "uses AI." Native.

What "AI-Native" Actually Means

There's a difference between using AI as a tool and building with AI as a partner.

Most companies that claim to use AI fall into the first category. They have a chatbot for customer support. They use Copilot for autocomplete. They feed their docs into an AI to generate FAQs. These are useful applications of AI, but they're add-ons—layers on top of how work already happens.

For us, AI isn't a layer. It's foundational.

"We want Ferni to remember things that matter to users, but not everything. How would a good friend decide what's worth remembering?"

The AI doesn't give us the answer. But it helps us ask better questions. And that's often more valuable.

The Human Decisions That Still Matter Most

Here's the thing nobody tells you about building with AI: it doesn't replace the hard decisions. It clarifies them.

What should Ferni care about? Memory systems could remember everything. We chose to remember what a good friend would remember—not obsessively logging every detail, but naturally holding onto what matters.
When should Ferni stay silent? AI can always fill space with more words. We chose to build in "meaningful silence"—moments where Ferni pauses, reflects, or simply waits because that's what a thoughtful person would do.
What won't Ferni do? The easiest features to add are often the worst for users. We've said no to features that would be "impressive" but wouldn't serve people well.

AI helps us move faster. But the compass is human.

Why We're Sharing This Openly

We could have quietly used AI in our development process and marketed Ferni as a purely human creation. Plenty of companies do this. We chose the opposite.

1. It's honest. Ferni's whole purpose is authentic connection. If we pretended AI wasn't part of how we work, we'd be starting our relationship with users based on a half-truth.

2. It's the future. The question isn't whether AI will be part of how software is built—it already is. The question is whether companies will develop good practices for working with AI, or stumble into bad ones.

3. We're proud of it. What we've built with AI assistance is better than what we could have built alone. Not because AI is smarter than us, but because the collaboration produces something neither could do separately.

What We've Learned So Far

AI accelerates iteration, not ideation. The big ideas—what Ferni should feel like, what problems it should solve, how it should treat users—those come from human insight and observation.

Conversation is underrated. We've found more breakthroughs talking to AI than talking at it. When we treat Claude like a thinking partner rather than a code generator, we get better outcomes.

Taste still matters. AI can generate infinite variations. Choosing which one is right—which response feels warm instead of performative, which architecture will scale gracefully—that's still irreducibly human.

If you want to see what all this building produces, try talking to Ferni. Web at app.ferni.ai, or call (888) 598-3952. Free to try. No signup required.