Stella Foster

“More life, less phone.” Few taglines pick a fight with the device they ship on. Stella Foster, from Santa’s AI Lab Inc., is an AI assistant for people who want to spend less time on a smartphone. It pursues that goal with an unusual distribution strategy. Alongside an Apple Watch app and a Google Play listing, the assistant answers an ordinary phone number. Anyone can call or text it from a flip phone, a feature phone or a landline. Callers can ask questions, get directions, set reminders, manage a schedule, follow news and sports, and send or read messages.

That choice gives Stella two audiences who rarely share a product. The first is the digital minimalist who wears an Apple Watch and wants the iPhone in a drawer; the watch app works with or without a paired iPhone. The second is the older adult for whom the touchscreen itself is the obstacle. Press coverage leans hard on that second group. KTLA’s report centers on the creator’s community classes, where older attendees said AI could help them.

The pitch goes beyond answering questions. The Google Play listing describes Stella as a personal digital robot that knows its user and takes action for them. It texts and calls on the user’s behalf, remembers contacts and preferences, and sends follow-ups. That moves the product from assistant into agent territory, which raises both its ceiling and its stakes.

What Stella lacks is a public record. Public reception is thin: the App Store has no meaningful rating overview, Google Play shows no meaningful rating, and a Show HN post drew little discussion. KTLA’s segment is warm, but it quotes only the creator and supportive attendees. There is no independent reception yet, so Stella has to be judged on its design.

• Phone-number access from any device: Users can call or text Stella from flip phones, feature phones and landlines. Nothing has to be installed on the caller’s handset.
• Standalone Apple Watch app: The app runs with or without a paired iPhone. It offers chat with image support, calling, account management and a watch face widget for one-tap calling.
• Delegated communication: Stella can make and receive calls, send and read messages, text and call on the user’s behalf, and send follow-ups.
• Persistent memory: Stella remembers contacts and preferences, so requests can stay short and conversational.
• Everyday assistant tasks: Questions, directions, reminders, scheduling, news and sports.
• Mac message bridge: A Mac desktop app connects iMessage, WhatsApp and Telegram. Those conversations can then be reached from a second phone through Stella’s number. The Mac app must stay open and running.
• Reported pricing: KTLA reported a free tier with a daily calling allowance and texting, plus a paid monthly subscription for more calling time.

A Phone Number as the Interface

The most consequential design decision in Stella is also the least glamorous: a phone number. Voice calls and SMS are the most universal interfaces in consumer telecom. By meeting users there, Stella avoids the onboarding friction that sinks most assistants aimed at older adults. The caller’s own handset needs no app store, no update prompts, no login screen and no gestures to learn. The mental model is “call Stella,” which anyone who has phoned a pharmacy already understands.

The trade-off is feedback. A screen shows what an assistant heard and what it is about to do; a phone call does not. For lookups such as directions, a score or a headline, a misheard request costs a few seconds. For actions, it costs more. An assistant that sends a text in someone’s name needs a confirmation step that is explicit without being tedious. Prospective users, and the families helping them set up, should check how Stella reads back an action before trusting it with a contact list.

Agent Ambitions, Trust Questions

Memory is what makes the agent pitch useful. An assistant that already knows who “my daughter” is turns a three-step request into one sentence. Follow-ups let it finish tasks a forgetful user might drop. For an older user who struggles with modern devices, delegation is the whole point.

Delegation is also the whole risk. Stella stores contacts and preferences, calls and texts in the user’s name, and can reach iMessage, WhatsApp and Telegram through the Mac app. Together, that is a large trust surface, run by a small developer with no public reception to vouch for it. None of this makes Stella unsafe. It does put the burden of proof on Santa’s AI Lab. Before we would recommend Stella for a vulnerable user, we would want clear documentation of how it stores data and when it acts without asking.

The Watch as Primary Device

On Apple Watch, Stella gets a native surface: chat, calls, account management and the watch face widget. That widget is the smartest small decision in the package. It turns the assistant into a speed-dial on the wrist, which suits a voice-first product better than a chat thread on a tiny screen. Image support in chat broadens what a wrist conversation can carry.

Running without a paired iPhone matters to the minimalist audience. It lets Stella be the reason the phone stays home, rather than one more app that needs the phone nearby.

The Mac Bridge: Clever and Fragile

The Mac desktop app is Stella’s most ambitious feature and its weakest link. It routes iMessage, WhatsApp and Telegram through Stella’s number to a second phone. That gives users a way to carry a basic handset without dropping out of their group chats. The catch, stated in the product description and the Show HN discussion, is that the Mac app must stay open and running. If the app quits or the Mac shuts down, the bridge goes dark.

That requirement cuts against the tagline. A product that promises less phone asks for an always-on computer instead, and only households with a Mac can use the bridge at all. For a tech-comfortable minimalist with a desktop that never sleeps, it is a smart workaround. For the older adults the press coverage targets, it is a poor fit. A hidden dependency on a computer in another room is exactly the fragile setup that ends in support calls to relatives.

Pricing Structure

KTLA’s reported structure is easy to understand: a free tier includes a calling allowance and texting, and a paid monthly subscription buys more calling time. The free allowance makes a trial low-risk, and text is the cheapest channel. The meter runs on voice, which is the channel landline and flip-phone users will rely on most. If calls placed through Stella draw from the same allowance, people who route their real conversations through it will reach the ceiling first. For them, the monthly subscription is a significant recurring cost. These terms come from a press report, so subscribers should confirm the current terms before paying.

For a large share of its intended users, Stella has no visual interface at all. A landline caller experiences it as a voice, and a feature-phone user experiences it as plain SMS. The design burden therefore falls on conversational clarity: pacing, read-backs and plain language matter more than art direction.

Where visuals do exist, the priorities make sense. The one-tap watch face widget reduces the interface to a single target. That is the right call for users who struggle with small screens or have limited dexterity. Fitting chat, images and account management onto a watch is harder to keep legible, and that is where the visual design will be tested.

On performance, Stella’s architecture shifts the work away from the user’s device. An assistant reachable from a landline has to do its processing remotely. As a result, a decades-old handset can reach the same assistant as a new watch, and hardware is not the constraint. Responsiveness depends on the network and Stella’s servers instead. Voice is an unforgiving medium for latency: a pause on a call can sound like a dropped line, while the same delay in a chat window goes unnoticed. The Mac bridge can only be as reliable as the Mac that keeps it running.

• iOS / Apple Watch: This is the most fully described native experience. The watch app runs with or without a paired iPhone and offers chat with images, calling, account management and the one-tap watch face widget. Input is touch and voice on the wrist.
• Android: The Google Play listing leans on the agent pitch: a digital robot that knows the user, acts for them, remembers contacts and preferences, and follows up. The wearable features in Stella’s descriptions belong to Apple Watch, so Android users should not assume an equivalent.
• Mac companion: The iMessage, WhatsApp and Telegram bridge requires a Mac that stays on, whichever phone the user carries. It suits households already invested in Apple hardware.
• Any phone: The call-and-text channel works the same on every device, from landline to flip phone. Input is voice, keypad and SMS, with no app involved.
• Pricing and reception: KTLA’s reported pricing is framed around calling time and texting, not platform. Public ratings are absent or negligible.

Stella Foster’s structural idea is strong: an assistant reachable through an ordinary phone number removes the smartphone from the interaction. That design serves two narrow audiences—watch-wearing minimalists and older adults who avoid touchscreens—and the one-tap watch widget and contact memory reduce steps.

But the product is hard to recommend for most people. An agent that speaks in its user’s name needs a track record, and Stella has almost none: no meaningful public ratings and only sympathetic press coverage. The Mac bridge undercuts the less-phone promise with an always-on computer, and the paid voice tier will matter to anyone using Stella as a real calling line. The free tier is the sensible starting point, especially for a relative who struggles with smartphones; independent reception should come before handing over a full contact list or a subscription.