Hello Aria - Turn chats into tasks, reminders & notes — instantly.

@sai_tharun_kakirala the 'MOM Mode' (meeting minutes) from voice notes sounds incredible for small teams. we spend more time writing summaries than we do in the actual meetings sometimes. rooting for you guys.

@sai_tharun_kakirala this is super smooth. Well done. gl with the launch 🫡

great! @sai_tharun_kakirala are you routing whatsapp messages through a llm api or running a custom model?

@davitausberlin 
Thanks Davit 🙏

Small but important distinction — HelloAria isn't a chat hub that aggregates conversations across apps. It's an AI assistant that lives inside the messengers you already use (WhatsApp, Telegram, email), and turns what you say into structured output — tasks, reminders, calendar events, meeting notes.

So instead of "one inbox for all chats," it's "one brain that understands what you need and executes it across tools."

On B2B — 100%, it's already a big use case. Outlook is live today, along with Google Calendar, Drive, Meet, and OneDrive. What's coming next is deeper team functionality — shared reminders, meeting minutes with auto-tagged action items per person, and proactive nudges like "hey, you promised the deck to Sarah by Thursday."

Slack + Teams integrations are on the roadmap too.

What's your biggest B2B coordination pain right now? Genuinely curious — we're actively shaping that roadmap.

Good luck with Banyan AI Lite 🚀

@julialeffler 
Oh this one hits 🙏

Parents were genuinely one of the use cases that surprised us most. We built Aria thinking "busy professionals" — and then a bunch of moms and dads wrote in saying "I finally added the dentist appointment because I could voice-note it while holding the baby." That reframed the whole thing for us.

If you try it: voice-note something like "remind me Thursday at 7am to pack Emma's water bottle" — Aria will parse the day, time, and your kid's name and just handle it. No typing, no switching apps.

Would love to hear how it holds up in real kid-chaos — that's the feedback that shapes the roadmap most 🙏

Thanks for being here on launch day ❤️

@osakasaul 
Thanks 🙏 and this is genuinely one of the hardest problems we've solved — love that you zeroed in on it.

Short answer: fully seamless, zero manual management.

How it works under the hood: the channels (WhatsApp, Telegram, iOS, email) are just interfaces. Aria's "brain" sits above all of them — a single user context with unified memory, task state, and history. You're not talking to "WhatsApp Aria" or "Telegram Aria," you're talking to Aria, who happens to be reachable through those channels.

So the flow you described works exactly as you'd hope:

→ Create a task on WhatsApp → "remind me to send Q3 report Friday"

→ Later on Telegram → "what did I need to send Friday?" → Aria recalls it instantly

→ Mark it done from the iOS app → all channels reflect the update in real time

The key architectural choice was making the channels stateless and the brain stateful. Most multi-channel tools do it the other way (each channel maintains its own state, then they try to sync) and it's why those products feel fragmented. We learned that the hard way in v1.

Biggest unsolved part we're still working on: cross-channel conversation continuity — picking up a half-finished conversation mid-thread when you switch devices. Much harder than state sync.

What made you ask this specifically — are you building something adjacent?

@saad_el_gueddari 
Thanks 🙏 and yeah — that's exactly the conversation we had internally about 8 months in.

HelloAria actually started as WhatsApp-first AI. It was getting traction, users loved it, but every time Meta shifted a Business API policy I was losing sleep. One policy change away from an entire user base being locked out wasn't a risk we were willing to ride long-term.

So we rebuilt the architecture to be channel-agnostic. Today Aria lives on:

→ WhatsApp

→ Telegram

→ iOS app (our own platform, fully controlled)

→ Email (SMTP — nobody can take this away)

→ Web dashboard

Same AI brain, same user context, same data — users pick the channel that fits their life. If Meta changes something tomorrow, our WhatsApp users migrate to Telegram or the iOS app in one tap and keep all their tasks, reminders, and history.

It cost us 3 months of engineering we didn't really have, but channel dependency is one risk I don't want to carry into a Series A conversation.

Good eye on this — most people don't spot it until it's already a problem 🫡

@sounak_bhattacharya 
Both great questions — and honestly the two that define whether voice-to-task actually works in the wild vs just demos well.

On latency:

A 30-second voice note with 10 items shows up as structured tasks in under 30 seconds after you hit send. The parking-lot scenario is genuinely the use case — fast capture, read it back later when you're at your desk.

On judgment calls:

Aria classifies without asking, 90% of the time. A few heuristics:

→ Explicit time = reminder ("remind me at 5pm")

→ Explicit date / no time = task ("buy milk tomorrow")

→ No time anchor + declarative statement = note ("thinking about restructuring the pricing page")

→ Person + action = follow-up ("tell Priya about the Q3 thing")

The ~10% edge cases — things like "think about the new hire" could be task OR note — she picks the most likely bucket and lets you override in one tap from the iOS app. We intentionally don't ask clarifying questions mid-voice-note because every back-and-forth kills the "parking lot speed" magic.

The worst user experience isn't wrong classification — it's interrupting the flow to ask. People forgive "you put this in notes instead of tasks" way more than "you made me answer 3 questions after a 30-sec voice memo."

What's driving the question — personally a voice-note heavy user, or thinking from a product angle?

@byalexai 
Thanks so much 🙏 and yeah, that's honestly one of my favorite questions to get asked because the answer surprised us.

Intuition said WhatsApp wins — it's where people live, lowest friction, instant habit.

Reality: WhatsApp wins for acquisition, but iOS wins for retention.

What we see: users onboard through WhatsApp (because that's where the friction is zero), then graduate to the iOS app within 2-3 weeks for anything involving multiple reminders at once, voice-note meeting recordings, or reviewing their day. WhatsApp stays as the "quick capture" layer — the "remind me to call mom" moment while walking.

So it's less "which channel wins" and more "which job-to-be-done lives on which channel":

→ WhatsApp/Telegram = capture (fast, one-shot, voice or text)

→ iOS = review + meeting recording (richer UI, full context)

→ Email = reports, summaries, anything you want to archive

The biggest surprise: power users end up using ALL of them, daily, without thinking about it. Aria syncs the context across, so they don't feel like separate tools.

What's your context for the question — building something channel-first yourself?

@byluocca 
Thank you so much — that’s actually a really thoughtful point.

You’re right, people who travel a lot or have friends and work contacts across different countries often end up juggling different messaging platforms depending on where they are. That cross-platform chaos is exactly the kind of friction we want HelloAria to reduce.

It may sound niche at first, but it’s a very real use case — especially for people managing personal and professional communication globally. Really appreciate you pointing that out.

@uxpinjack 
Thanks 🙏 and yeah, "tool overload" is exactly the framing we keep coming back to — people don't need another app, they need fewer.

On scale: the short answer is Aria treats memory like a human assistant would, not like a database. Three layers working together:

→ Short-term context — what you're talking about right now (this week's meetings, today's to-dos)

→ Mid-term memory — recurring patterns Aria picks up (your Monday 9am standup, the grocery list you rebuild every Sunday, your kid's swim class)

→ Long-term knowledge — the stable stuff (your family's names, your work projects, recurring contacts)

Retrieval is semantic, not chronological — so when you say "remind me about that thing Sarah mentioned," Aria searches intent, not keywords.

The honest challenge at scale isn't storage, it's forgetting gracefully — knowing when an old reminder is no longer relevant vs when to surface it again. Still actively tuning that one.

Great question — what kind of usage were you imagining when you asked? Teams, personal, something else?

@tommy_clarke 
Oh the UK WhatsApp angle is SO real — genuinely one of our top user clusters. Most people's social lives in the UK happen in WhatsApp groups (pub plans, 5-a-side footie, birthday drinks, kids' school coordination) and it's mad that none of that flows into a calendar by default.

On your two questions:

1. Forwarding group messages to Aria:

Forward any WhatsApp message to Aria (or copy-paste it into your chat with her) and say "make this a calendar event" or "remind me the morning of this." Aria reads the message, pulls out the date/time/location/people, and creates the event. Works great for those "drinks Saturday 7pm at the Red Lion" messages that normally get lost in a group thread.

2. Memory + recall:

Yes — today Aria has full context of your upcoming events and reminders, so "what's on this week?" or "what do I have on Saturday?" works instantly. She can also pull from your notes when you ask — "did I note anything about the dentist's address?" kind of questions.

One small thing forgetful folks love: voice notes. If you're walking home and suddenly remember 5 things, just voice-note them to Aria — she'll structure everything into tasks/reminders without you needing to type a word.

If you try it, would genuinely love to know how it fits into a UK social-WhatsApp life — that's exactly the use case we want to nail down further 🙏

(And we've got 3 months free for the PH community — details in the maker comment pinned at the top 🚀)