This AI-Powered App Turns Family Memories Into Personal Podcasts

What if AI could help you preserve your grandfather’s stories, or even create a podcast out of them? That’s the promise of Linda AI, a new startup turning memories into shareable audio shows with a conversational AI as your guide. Users can use the platform to have a professional interviewer-like bot probe a subject with questions and then convert those verbal responses into an engaging audio program. Although recording the memories of yourself and loved ones is an ideal use case for a tool like this, it’s easy to see its application in other scenarios.

“My goal with Linda was to create the easiest way to pull out incredible stories from people so they can share them with family, friends, or the whole world,” Nick Hasty, Linda’s founder and the former founding engineer at GIPHY, tells me in an email. “Talking is the most natural way to do this, and the podcast is then the perfect format for sharing. What’s missing is the podcast host who will ask the right questions and create an amazing conversation worth listening to.”

Helping Older Users Create Content

According to The Atlanta Journal-Constitution, the company’s name comes from Hasty’s mother after the two swapped stories while on a road trip a few years ago. In fact, families are the first group of users that Linda is trying to reach, as they’re “so dispersed now and people don’t spend quality time together like they used to.”

More specifically, Hasty believes there’s an opportunity in going after older adults. He says that by 2030, there will be more people over the age of 55 than there are children. “This demographic is grossly underserved with regards to having apps and software designed to help them create content.”

Hasty continues, “Less than 7 percent of adults over 65 actively create online content, compared to over 70 percent who consume it. But those who do create content report better mental well-being and competence levels. So I wanted to build something for this demographic of people, who have already led rich lives and likely have a few decades ahead of them yet, enabling them to create a living archive that would grow and evolve with them.”

Nevertheless, the company has a wide range of users spanning demographics, all of whom want to talk to an AI and create podcasts. So while Linda is primarily reaching older generations, it’s developing new capabilities to help users create content for a wider set of use cases beyond the “life story” angle.

An AI-Powered Content Guide

Using Linda is like a reverse prompt—instead of generating a prompt, the human selects the prompt (“topic”) to answer. The subject provides the answer verbally to all related questions before it’s all turned into a podcast episode of the conversation. Each one can be up to 15 minutes in length. If you want to share the files publicly, they can be published on your Linda account page, Spotify, or Apple Music.

“There’s a universe of incredible ‘content’ locked inside people’s heads…unique perspectives on life, hilarious anecdotes, singular wisdom,” Hasty explains. “The vast majority of these experiences will never be written down, much less put on the internet. I believe most people need a little help opening up, however, so Linda is designed to ask people amazing, personalized questions and get them talking.”

The service has two subscription tiers: a free version that allows access to two interviews and podcast episode creation and hosting, and a $9.99 monthly “legacy” plan that provides unlimited access to interviews, custom interviews, personalized conversations, automatic podcast episode creation, and data downloading.

What users see on Linda may seem straightforward, but behind the scenes, it’s what’s described as being “quite sophisticated.” The company uses state-of-the-art speech recognition and advanced large language models (LLMs) capable of understanding context and generating relevant follow-up questions. These same models are also used to create the podcast episodes. Linda is also leveraging specialized models to polish the audio, removing background noise, generating music for the intro and outro parts of the episode, and more.

“What makes Linda special is how all these technologies work together behind the scenes while the user just needs to talk,” Hasty proclaims.

All this being said, Linda isn’t the only tool to record or produce an audio show. Alternatives include Descript, Squadcast, Wondercraft, Zencastr, and Google’s NotebookLM. However, perhaps the differentiator that Linda currently offers is its AI-powered interviewing technique, whereby the bot listens to responses before crafting follow-up questions.

“I like to think of Linda as NotebookLM for people instead of PDFs,” Hasty says. “Where NotebookLM has two hosts that talk about the contents of documents, Linda is the sole host, and the user is the source of content. Linda’s job is to ‘query’ the user and get them to open up and share the stories and insights they have. What’s fascinating is that Linda inverts the normal relationship between user and AI—instead of you prompting the AI, Linda has to figure out the correct prompts/questions to traverse your brain’s ‘latent space’ and bring out the most compelling content. Unlike most AI tools, where you need to know what to ask for, Linda guides the entire experience so you can kind of relax a little and focus on your memories and feelings.”

Data Security

As with any AI product, the risk of hallucinations is a top concern, especially when using a tool like Linda to capture something as personal and meaningful as family memories. The company acknowledges this fear, but claims it has installed “solid” guardrails to keep the LLM in line.” It shares that when addressing core interview topics that delve into love and childhood, Linda uses pre-designed scripts to inform it of follow-up questions. In addition, because the app has access to the running transcript and key facts about the user, there shouldn’t be any shortage of context to reference. Hasty states that when there has been any confusion in the conversation, “it’s usually from Linda misunderstanding what the user has said, just like when talking to a human.”

When asked about data security, Hasty responds like most other companies, remarking that his startup has a privacy-first approach and users are in control over episode visibility. Audio files are also stored in private cloud storage, isolated away from the internet. Private episodes are separated from published ones and feature a clear consent warning. Moreover, Linda blocks AI bots and web crawlers from indexing users’ content.

‘The YouTube of Personal Podcasting’

As mentioned earlier, Linda is working to expand its appeal beyond the “life story.” The app could be helpful to job seekers looking to promote their work experience or identify thought leadership opportunities. Small businesses could leverage it to craft customer testimonials. Lastly, younger audiences might use it to help them through their therapy. These are all examples Hasty believes could be possible.

“I want Linda to evolve into a platform that helps users create whatever kinds of audio content they want, with multiple use cases,” he emphasizes. “So you’ll tell Linda who you are and what kind of podcasts you want to make, and she’ll adapt her conversational style, interview topics, and podcast episode format to be more in line with your individual goals.”

But why audio over video? Hasty claims it’s because there’s a different kind of connection. People tend to dig deeper and have more meaningful connections with audio programs. He states that voice has more authenticity than images and videos. It also forces people to use their imagination more.

Ultimately, Linda strives to be the “YouTube of personal podcasting.” The company will eventually add creator-centric analytics, improve content discoverability, and build radio-like channels around curated themes and topics.

To date, Linda has received angel funding from Factorial Capital and is in the process of raising a seed round that will be used to build out a larger personal podcasting network. It has a lean team of at least four employees, which Hasty states makes the company more nimble and can adapt quickly to user feedback.

Linda is currently only available on the iPhone.

Featured Image: An AI-generated image showing an older person recording a podcast. Credit: Adobe Firefly