PH热榜 | 2026-07-30

Hey Product Hunt 👋 I'm Anand, one of the makers.

We build with coding agents every day, and kept catching ourselves typing a three-sentence prompt we could've said out loud in four seconds. Thinking is fast. Typing is not.

So we built SKI. You hold the Function key, say what you want, and your agent goes and does it — then answers you out loud, like a real person. It's an ambient thing: it sits in your notch and never interrupts your flow.

Three things we care about:

• It answers. Voice input for agents already exists — it's one-directional, words in, text out. SKI closes the loop: it speaks the result back.

• It comes to meetings. Bring your coding session into a call as a real teammate and build live — or send your agent in your place to speak for you.

• It's local. Speech in and voice out both run on your machine. No cloud, works offline.

Free for life — Mac and Windows today, Linux next.

First launch for us, so hard questions are genuinely welcome. What agent do you use, and what would make you keep this open all day?

Does it have built in intelligence to understand and respond ? Or is it working like bring your own agent where the speech understood and spoken by Claude Code or our agents ?

Cannot wait to try this out 😀. I’ve been trying to build this quite a few times myself. The models just recently got good enough to be able to pull this off.

Can you also share your screen with the ai agent? It’s nice to be able to show it a problem

This is genuinely interesting. Feels less like voice dictation and more like collaborating with an AI teammate. Congrats on the launch!

Will it run on agents running on isolated PCs without internet connection ?

The meeting use case is interesting. Bringing an agent into the conversation instead of just using it behind the scenes is a cool idea. Excited to see where this goes.

The voice loop makes sense for coding, especially when the prompt in your head is much faster to say than type. How well does SKI handle noisy rooms or shared offices? I can imagine fans, keyboards, background conversations, and meeting audio making this tricky pretty quickly. Does it have any kind of voice isolation or push-to-talk filtering, or is it mainly designed for quieter environments?

Dictation and voice commanding fail differently, and I think that is the interesting part here. A misheard word in dictation sits on the screen and you fix it. A misheard word here goes to something that acts.

Delete the test file and delete the rest of the file are both fluent, both plausible, and ASR will be confident about the wrong one. So confidence scores do not save you, and push to talk only controls when it listens, not what it commits to.

You speak the result back, which closes the loop after the agent has already done it. Does it ever repeat the instruction before, on the ones that are hard to undo?

Very promising approach - two things which made me uninstall it after a few minutes:

a) I need a way to select input and output. I have multiple mics (Jabra wireless headset for meetings, Rode mic for video recordings, some system loops etc) and also multiple outputs. (When it gets late I want that output on my headset to not annoy my family)

b) price intransparency. I understand that you need to make money. We all do. But if you tell me "free for lifetime" then I am annoyed if I see $2.50 somewhere in the fineprint of my account status. Be open about your pricing. You are competing against other tools that I pay for and I am happy to pay for a better tool

Is the offline meeting recorder compatible with all meeting platforms?

I am a firm believer of the rapid the migration from traditional UX to voice based UX. Would love to check this out and give feedback. Kudos on the launch, SKI team!

thats interesting, using voice to talk to ai and code is an amazng idea but do people actually want to talk or just type???

@johnkg003 Congratulations. And happy product launch.

Quick video on how to use SKI https://www.youtube.com/watch?v=8bE4L6JAQ-o

Congrats on your launch! Would like to see Linux support!

Hi,
What languages are supported?

This is interesting. How do you highlight a specific code in a file using dictation? Just curious :)

Small feedback upfront, when I simple click on "Download for Mac" on the website, the download on my device doesn`t start. Workaround found, if I do right click and open in a new tab, then it start the download.

Congrats on the launch. Just a quick question, how does the push-to-talk handle noisy environments or mechanical keyboard chatter during active coding?

This is definitely a valuable addition to the coding community. I currently use whisper flow for coding but I am wondering how this differs and maybe how it will be better given that it's tailored toward programmers.

I run Claude Code all day as a solo dev and the thing that actually breaks my flow is drifting to another tab and not noticing when it's done or sitting blocked on an approval. Voice-out closing that loop is the part I actually want — does it speak up when the agent is waiting on me, or only when a task finishes?

Congrats on the launch. Assuming this can work with local models since you said it works with Hermes?

SKI helps you get the most out of your Claude Code or Codex subscription.

Beyond interacting with your AI agent, you can also use SKI for note-taking and organizing your ideas.

https://www.youtube.com/watch?v=6VTptRCOA_w

Just curious... Is notch mode available in windows ?

In windows, it just overlays on top. I’d rather be able to hide it when I don’t need it. Notch seems like a good solution. Is there anything similar available for Windows?

Congrats on the launch! Voice input for coding agents makes sense in a way it never did for typing code, since you're dictating intent rather than syntax. Curious how you handle the context switch when you need to reference a specific file or function mid-sentence. Is there a way to point at things while talking?

Congrats on the launch! Voice input for coding agents feels overdue. Does it handle multi step commands, or is it mostly one shot prompts right now?

The teammate framing is fun, having Claude Code answer me out loud feels closer to pairing than prompting. I could see myself using this while pacing around the room thinking through a refactor. Does it hold up in a noisy space like a shared office, or does it really want a quiet room?

Hey Product Hunters 👋

I'm Zhiqi, one of the makers behind Memmy.

If you bounce between Cursor, Claude Code, Codex, and other AI tools like me, you know the pain: every new chat starts from zero. Prefs, decisions, project context — you HAVE TO re-explain to them again and again.

Finally we get rid of it, and we're excited to introduce Memmy 🍙

Memmy is a local-first memory layer those tools can share. Auth once, and it turns messy chat history into structured memory — prefs, decisions, project context. Switch tools, and Memmy carries what matters with you. No re-explaining anymore. You can search it, control who uses it, or delete anytime. And it stays on your device, never uploaded to a remote server.

Also an agent itself: you can hand it tasks, not just manage memories.

We're open source, free to start, and would love to hear your thought — "What tool should we support next, and what should memory never forget?"

We gonna be here all day to answer questions and would genuinely appreciate your feedback 🫶🏻

Been waiting for something like this, honestly.
BTW, does it work well with Claude Code specifically, or is that still early days?

Love the pitch! How do you handle conflict resolution when a user exhibits different personalities or prompt styles on different tools? (e.g., coding on Cursor vs. creative writing on Claude)?

Memmy is a long-term memory assistant for individuals and developers, built to break down the memory silos created when working across AI coding agents such as Cursor, Claude Code, and Codex.

It consolidates historical conversations, task progress, and key decisions from multiple agents. Powered by the memory capabilities of MemOS, Memmy turns fragmented context into persistent, searchable, updatable, and reusable long-term memory. This means that when you switch tools, start a new session, or hand work off to another agent, your project context does not have to start from scratch.

Memmy is especially useful for:

• Developers and teams using multiple coding agents;

• Engineers who want to preserve architectural decisions, technical investigations, and debugging history;

• Users who want AI to continuously understand their preferences, workflows, and long-running goals;

• Teams exploring long-running agents, multi-agent collaboration, and AI-native development workflows.

In one sentence: Memmy helps AI go beyond remembering a single conversation—it helps AI continue your work with persistent context.

GitHub: https://github.com/MemTensor/memmy-agent

Congrats on the launch! We're building humalike.ai and personal memory is one thing we care about :))

Quick video on how Memmy works

https://youtu.be/QURd0wkf4gI?si=cOLHwrsSlWWi5wXn

congrats! @brattyaiguy

Congrats on the launch. Local-first is why I clicked, most memory layers ship as someone else's cloud. One thing I could not work out from the description: when Cursor and Claude Code both hit the OpenAI-compatible API, does a memory carry any scope with it, per project or per tool, or does every connected agent read from the same pool? I have work repos and personal side projects on the same machine and I would want the second not to bleed into the first.

Are decisions, preferences, project state, and lessons stored as different memory types with separate update rules?

Does the OpenAI-compatible API provide a way to inspect the memories used for each response?

oh this is a very clever one. AI not remembering anything about me when i switch (even between claude web and claude code, for example,.. same company...) is one of the most frustrating things I've dealt with.

Do you also have any ideas on how to make memory "better", i.e. have it be more comprehensive?

This solves one of the most frustrating parts of using multiple AI tools: having to explain the same preferences and project context over and over again. A shared memory layer feels like a missing piece in the AI stack. Congrats on the launch!

Amazing product! Congrats for this launch!

Local SQLite storage plus user-controlled access makes the privacy story easy to understand. I appreciate the concrete implementation details.

Local note apps preserve information, but they do not automatically surface the right decision to the agent working on today's task.

Retrieval analytics showing which memories helped, were ignored, or caused corrections could improve both user control and system quality.

I’d use this to keep my writing voice consistent across different AI tools. They all seem to understand my style eventually… right before I start a new chat

Referencing API keys through environment variables instead of hardcoding them in config is exactly the kind of secure default I want.

Been looking for something like this, using claude_mem skill up to now.
Amazing!

Could definitely see this being used for shopping or booking flights. I want my agent to just remember my preferences and what I like in terms of clothes and flight selections for example.

Interesting thought: could a team use Memmy to help onboard new hires? There’s so much useful context that never makes it into the official docs.

This could save people a lot of time. Congrats, team!

local-first with no cloud is exactly why i clicked too, but what happens on a new machine? does the memory pool migrate somehow or does a laptop swap just reset you to zero. that tradeoff between local-only and continuity feels like the real design fork here

Memory with agentic flows is often a gap when I’ve tried my own experiments, looking forward to checking this out!

This is really cool! Now I just need to figure out who I am.

Memory staying local is pretty key. Great to see it!

This is awesome. I think the whole community been waiting for something like this.

BTW does it work with any AI tool already??? Like Antigravity, Grok build, etc...

I work across several products on one machine and a couple of them handle clinical data, so my question is about isolation: can memory be hard-scoped per project, or is separation just an outcome of retrieval ranking?

Re-explaining the same context to every new AI tool is quietly one of the most annoying parts of this whole space, so shared memory across Claude Code and Cursor sounds like a small dream. Open source makes it easier to trust with this kind of data too. If I open my memory up, what does it actually look like, plain files I can edit or something more structured?

Interesting idea, you have my upvote. What are the current limitations?

Hey Product Hunt! 👋

I’m Ilia, founder of AI Search Console.

We built it after repeatedly hearing the same question from SEO agencies and their clients:

“Are we actually showing up in ChatGPT?”

The usual way to answer was surprisingly manual: run a few prompts, take screenshots, copy the results into a spreadsheet, and try to guess whether visibility was improving.

That approach breaks down quickly. AI answers vary by prompt, model, market, and time. A handful of manual checks cannot reliably show your share of voice, explain why a competitor appears more often, or identify which sources influence the answers.

AI Search Console turns that process into repeatable analytics. It monitors how your brand and competitors appear across ChatGPT, Claude, Gemini, and Perplexity, down to the individual prompt and cited source.

What you can analyze:

– Prompt-level brand mentions and rankings
– Multi-model share of voice
– Competitor visibility and content gaps
– The domains and pages cited in AI answers
– The prompts where your brand appears and where it is missing
– Trends over time and client-ready reports

This matters because AI visibility is not determined only by your website or Google rankings. AI platforms may rely on review sites, editorial articles, communities, comparison pages, and other third-party sources. You can rank well in traditional search and still be mostly absent from AI-generated recommendations.

AI Search Console helps teams move from:

“Let’s manually check a few prompts”

to:

“Here is our visibility, here is the gap, and here are the sources and prompts behind it.”

It’s built primarily for SEO/GEO agencies and brands that need a measurable way to understand and improve their presence in AI search.

I’ll be here all day and would genuinely love your feedback. Ask me anything 🙏

💜 You can also try our demo: https://app.search-console.ai/demo

Hi! Congratulations on the launch!

How long will the free trial last? And what happens after it ends?

Do you recommend prompts during onboarding?

Do you track logged-in model experiences?

Looks nice, do you support Google Sheets exports?

Do you support role-based user permissions?

Congrats!
Do reports support custom branding?

@gorns Congrats! Do you track model version changes?

Congrats on the launch, and good luck!

Wow, the integration of search console and the poer of AI is what we marketers were needing.

Congratulations on the launch!

It's really interesting for me if results are stable. I mean if my and my friend's results in AI search will be the same in all the models

Congrats on the launch!
What happens if our prompt list is poor?
And what are your free trial terms? Couldn't find this in the FAQ

Can we import existing prompt lists?

Best luck, guys! Which AI models do you currently track?

How do you select relevant prompts?

Hi, great product. Beautiful, pleasant design, and a good name.

The only thing that's unclear is the pricing - what a "credit" actually is and how they're counted.

cited sources is the field i actually care about. i publish technical posts and i can see reactions and views, which tells me almost nothing, because a reader who got the answer from an assistant never shows up in any of my numbers at all.

the one i would gently push back on is ranking and share of voice. those come from a world with ten positions on a page. in an answer there is no position, you are in it or you are not, so cited or not cited is the real metric and a rank number risks being the comfort version of it...

Can prompts be assigned to different markets?

Can reports be exported as PDF?

Good luck guys! Congrats on your first launch btw

Is there a free trial or demo workspace?

Cool! How do you calculate share of voice?

Hey
Do reports support custom branding?

Great update! Congratulations!

Presence and correctness are two different problems and only one of them is in these metrics. Being cited says you were in the answer. It does not say the answer described you correctly.

The version that costs money is a confident wrong attribute. The answer says you integrate with something you do not, or that there is a free tier when there is not. The reader never opens the source, so they arrive already believing it, and share of voice scores that as a good day.

Do you capture what the answers claim about a brand, or only whether it appeared?

We looked at building a scrappy version of this in-house a few weeks ago and stopped, for a reason I'd be curious how you handle. At zero citations there is nothing to measure. Our site is young enough that every prompt came back with no mention, which is a valid data point exactly once.

So the tool seems to earn its keep somewhere above a zero baseline. Do you have a sense of where that floor sits, or do early-stage users get value from the competitor and cited-source side while their own mentions are still empty?

Good luck

Congrats on the launch! Finally someone's bringing analytics sanity to the wild world of AI search citations. Here's to fewer mystery prompts and more clarity!

Really exciting launch—making AI-search visibility measurable is a great idea. Congrats to the AI Search Console team! 🚀

Hey Product Hunt, Manouk here, founder of LangWatch.

We built this because we run a fleet of coding agents ourselves and had no idea what they actually cost: subscriptions hide the number, and local trackers only see one machine.

You know what Claude Code costs you per seat. What you probably can't answer is which sessions burned the budget, which model did the work, or whether your cache was earning its keep. Today's token prices are heavily subsidised, and the teams who find out what they actually consume after prices move are the ones who get surprised.

This week we've added a new launch in LangWatch, which our dev-team is using day-in, day-out. You can now track your full Claude Code usage on LangWatch: every session, tokens spent throughout the month, cache hit analysis, bash commands, skills and MCP tool calls, with a full reproduction of your terminal right in the UI.

Getting started is one command:

npx langwatch claude 

From there, every session lands in LangWatch automatically:

Where the tokens go: cost per session and per model, with cache reads and writes accounted separately.

Theoretical vs billed: on a Max plan, see what your usage would have cost at API prices.

Every tool call: bash commands, file edits, skills, and MCP calls, each as a span with duration.

Terminal replay: step through the whole session in the UI, as it happened.

Works the same for Codex, Gemini CLI, and opencode.

Free for individual use.

Start tracking your own sessions

I'll be here all day, happy to answer anything, and genuinely curious what you'd want to see next in it.

Congrats! How much overhead does the tracking introduce during long-running Claude Code sessions?

Good to note that individual developers can use it for free before rolling it out to larger teams. The pricing that I see on the top nav is for LangWatch or for this specific tool?

Manouk's "we run a fleet of coding agents ourselves and had no idea what they actually cost" is the honest version of this. Same here — our answer was "we'll find out when the subsidies stop."

One push: cost per session can mislead. The number that changes how you route work is cost per outcome — per merged PR, per green CI run, per fix that actually held. A run that spends 3x on the expensive model and lands it first try beats a cheap model looping four times — but a per-session view shows the opposite, and quietly trains you to downgrade the model that was earning its price.

You're already logging tool calls as spansm that: span → the commit or test run itproduced. On the roadmap, or are you deliberately staying at the session boundary?

+1 to the immutable price-stamp point above, with a wrinkle: it has to cut both ways. Rates must never change historical reports by accident — but "what the new prices" is exactly the question the subsidy argument makes people ask. That's a deliberate re-price, not a mutation.

I'm bookmarking this. Comparing theoretical cost against what I'm actually billed on Max is something I've wanted for a while.

This feels like the equivalent of DevTools, but for AI-assisted development. Congratulations on the launch @manouk_dr

Most observability tools focus on applications. This one focuses on the developers building them. Good to have LangWatch back on PH.

Theoretical vs billed is the number that decides whether you stay on Max or move to the API, so I'd put that one on the front of the dashboard rather than three clicks in. The part that will bite you is the price table itself. If you reprice historical sessions when a model's cache read rate changes, last month's report quietly changes too, and that's a number people build budgets on. Stamp each session with the rates that were in force when it ran.

Congrats on the launch!
Token tracking is becoming crutial with everyday usage of Claude... maybe eventually it will turn out that junior developers are less expensive after all :)

P.S:
I have built an interactive livedemo for you, feel free to check it out
https://app.livedemo.ai/livedemos/6a6afa27cc86a442a8fd9d17

does the terminal replay work for headless/CI runs with no TTY, or does capturing the replay need an actual interactive session? we run a chunk of our agent work in CI and cost visibility there is usually the blind spot

congrats on the launch! Does this use any bandwidth during the session? Especially long ones

Congrats on the launch! Seeing which Claude Code session actually burned the budget instead of guessing would be a relief, and the cache reads split out as their own token class is a thoughtful detail. How would this work for a small team, can we see everyone's sessions in one shared view or is it one workspace per developer?

The one-command setup makes trying it out much less intimidating. 😊

Congrats on shipping this! Finally a way to know exactly how much those late-night Claude Code sessions are costing me. My wallet thanks you already!

Can the platform can identify inefficient prompting patterns across multiple sessions?

the MCP tool calls as spans is the bit i want most. i drive a browser automation MCP from claude code all day and the screenshots are obviously the expensive part, except obviously is doing a lot of work in that sentence because i have no per-call number for any of it.

cache reads and writes counted separately is the other one, that is where the surprise usually hides.

really proud of this launch!

great product! knowing the per-seat cost but not which sessions burned is a proper problem, especially as the costs keep rising. this is going to save me from getting broke!

This is such a killer feature for heavy users and big teams with distributed sessions leveraging multiple providers! Great work, @LangWatch !

Knowing exactly what each Claude Code session costs is a game changer. Congrats on the launch, this is awesome! 🚀

The combination of cost tracking, session replay, and tool tracing makes this much more than another analytics dashboard. Congrats on launching!!!

This solves a problem many developers don't realize they have until the AI bill shows up. And if they have setup the credit card that's a red zone LOL

Looking forward to seeing historical trends showing how coding workflows evolve over months. :D

The replay functionality could be just as useful for learning as it is for troubleshooting. :)

I can imagine engineering managers using this to better understand AI adoption across teams. Good launch!!

I used to work as a Product Manager. Half my week went to the same questions: "How do I do this?" "Where is that?"

We did what every SaaS team does - videos, documentation, a help center, Chatbot. Users still asked. Videos cannot answer back and show them in real time.

NINA is the fix we wanted. It lives inside your product and answers users where they get stuck, by voice or text, guiding them step by step on the your product UI in real time.

And when you update your product, NINA keeps its guidance current. No re-recording walkthroughs after every release.

Setup is honest, not magic:

• Install the SDK and match NINA to your brand (about 10 minutes),

• Upload your docs (answers start immediately),

• Record your top workflows

The difference from your Looms: a video cannot answer back. NINA can.

Built for B2B SaaS teams running onboarding calls, supporting users in Slack or Discord, and maintaining tutorials nobody watches twice.

NINA is built by our team at AgenQ. Ask me anything I'll be here all day.

How do you get the context for the app? Do you feed some initial data and also monitor for changes?
Does your Agent hallucinate?

Congratulations on the launch! One thing I'm curious about: how does NINA handle situations where the UI changes frequently, such as products that ship updates every week? Does it automatically adapt, or does the team need to retrain workflows?

This is exactly the bottleneck most new SaaS products keep running into. Onboarding is still our biggest time sink. Docs and tutorial videos only get you so far when someone's actually stuck mid-flow and just wants an answer right there.

Congrats on the launch, looking forward to seeing how far this can take our onboarding.

Two questions:

• Any plans for Google sign-in, or is it staying limited to Slack/Microsoft Teams for now?

• Can we swap the narrator voice for our own, so it matches our brand instead of sounding generic?

Congratulations 🎊 👏 💐 🥳

Congrats on the launch!
Guiding users inside the application is crucial for them to adopt it.

P.S:
I have built a livedemo for NINA, feel free to check it out
https://app.livedemo.ai/livedemos/6a6afb52cc86a442a8fd9d95

It's an honor to hunt NINA here on PH, because I truly believe in its value proposition.

Many powerful softwares have no clear first starting points.
And NINA doesn't answer with steps. She walks users through them.

NINA guides users through real workflows inside your software, no training calls, no videos, no support tickets.

If resonates, help them by supporting their launch!!!

This is such a clever way to guide users right where they get stuck. Huge congrats on the launch! 🚀🚀

Congratulations on the launch! I wanted to know what languages does NINA support?

the voice side is what i want to know more about - is the TTS running locally or a cloud round trip, and does that add noticeable lag when someone's mid-flow waiting for the next step? that gap is the difference between feeling like a live guide and a chatbot with a delay

since voice input works too, does it pick up on tone at all, like frustrated vs just curious, or is it purely speech-to-text into the same guidance flow either way?

This looks super exciting - congrats on the launch! Real-time in-product guidance is exactly what users need.

Hi @varun1jan

Great product, when I logged in. To hear Nina was something else. But I got stuck while creating the workflow, asked Nina but I guess I already started the process and NINA could not help much.

Do you have some instructions there?

As someone just starting to build products, this looks really useful. Just in my testing so far I have seen the amount repetition that comes from user questions.

Congrats congrats! Just wanted to check if NINA works with both web and desktop applications? Kudos again!

The part I would ask about is the handoff. When NINA cannot finish something, whoever picks it up next needs the transcript and the steps already tried, otherwise the person types it all again and the deflection becomes a tax on the customer.

The other half is which questions guidance can reach at all. How do I is a UI question and you can walk someone through it. Why did this happen to my order is a record question, and pointing at the screen does not answer it. Deflection rate looks great on the first kind and never touches the second.

When NINA gives up, what does support receive?

Congrats on the launch! Does the assistant mostly work in a chat-like manner, guiding you through, or can it act as a fallback to a more interactive onboarding?

NINA looks genuinely exciting. Guide users step by step inside your product. is a strong problem to tackle—congrats on the launch! 🚀

damm, this is the right use of AI inside product we all wanted. Congratulations on the launch.

Congrats on the launch, Varun and team! The gap between documentation and helping someone at the exact moment they’re stuck is very real.

I especially like that NINA keeps guidance aligned as the product changes. How do you evaluate whether its guidance actually helped a user complete the workflow, beyond reducing support tickets?

Mohsin is asking the load bearing question and it has not been answered yet.

Selector anchoring is the most fragile thing you can build on. I write browser automation against other people's interfaces and class names churn constantly, sometimes inside a single session. A quarterly redesign is not really the threat. Tuesday is.

What has held up for me is anchoring on semantics rather than structure. Accessible name plus role survives a restyle, because a redesign usually keeps the button called Save even when every class around it changes. CSS classes are decoration, and they get treated like decoration.

If you want it genuinely durable, the honest version is a contract rather than magic: ask customers to put stable test ids on the elements NINA walks through. Less impressive on a demo, far less likely to break silently in production six weeks later.

What does it anchor on today?

@varun1jan Congratulations. And happy product launch.

Hey 👋, we’re Haz and Wyatt, the founders of Pally.

Some of you might remember us… a year ago we launched Pally as a personal CRM. It went well, but after thousands of conversations with users we kept hearing the same thing: nobody wants to manage their relationships in another tool. They want the stuff behind them (the replies, the plans, the follow-ups) to just get done.

So we rebuilt Pally from the ground up as a personal assistant that lives in your texts.

## How does it work?

There’s no app to open. You just text Pally like you’d text a friend (over iMessage or RCS), and it connects to everything else: your inboxes, calendar, files, tasks and the web.

Pally can:

- Reply to your unread DMs for you, in your tone (we’re the only text agent that natively connects to your iMessage and WhatsApp inboxes)
- Make real phone calls: book tables, chase quotes, sit through hold music so you don’t have to
- Book flights, buy things, and pay with locked one-time virtual cards
- Summarize the group chats you’ve been ignoring
- Remind you of things when you arrive somewhere (or leave)
- Keep track of every promise you’ve made, so you never leave anyone hanging (the CRM DNA lives on, just without the admin)

And it’s proactive. Pally texts you first: prepping you for meetings, flagging what needs you, and catching the stuff you’d forget (it’ll still remind you to ring your mum from time to time 🫡).

## Who’s it for?

Anyone who’s drowning in messages: founders, salespeople, recruiters, creators… but also just anyone who leaves people on read and feels bad about it (so, everyone).

## What’s next?

We’re launching publicly today, and we’ll be building in the open from here: more integrations (you can already bring your own via MCP), smarter proactivity, and letting Pally take entire workflows off your plate end-to-end.

Use this sign up link for a free month of Pally Pro: https://my.pally.com/r/GIFTPRODUCTH-FKFCSR

Come say hi in the comments, we’ll be here all day.

Thanks for the support,
Haz + Wyatt 🫶

I have a backlog of friends I ghosted on Whatsapp bcz I forgot to text back and now it's been so long and it's embarrassing. Hoping to power through them with Pally.

This product is 🔥 frictionless lifemaxxing fr. I'm the worst messanger of all time and this helps me to not let ppl down. Thanks pally team!

I want it to read Whatsapp voice notes!

Congrats on the launch! Can I send it voice notes?

Great stuff! Love having the interface in my messages, makes life so much easier all round.

Congrats on the launch looks sick! Curious, does Pally have access to Slack for context?

awesome tool - are you planning to launch a discord plug in ?

Congrats on the relaunch. I read the tone-matching line twice, because whatever writes in my voice has to have read a lot of me first. Where does that reading happen, on the phone, or on your side? My group chats are the last thing I would hand over without knowing that.

A happy Pally user here. Thanks for building this & congrats on the launch! @hazhubble @wylans

Really excited for this! How would you compare this product to Poke?

Congrats on the launch!

Love the deep integrations across everything + proactiveness. Keen to see where this goes.

does pally have an identity when talking to friends/groupchats? or does it respond as you?

Love the product! I've been a happy user! Congrats team!!

This looks crazy, wish I could try this on Whatsapp

Congrats !
You're plugging into iMessage & WhatsApp. What's the story on privacy ? Where does the data live ? and what does Pally actually get to see?
pleasure

Soo cool!! Needed this!! How do the auto reply settings work? For more predictable convos the auto reply feature seems neat!

Absolutely needed something like this, 1 day in and it has already helped me not feel stressed about my Whatsapp and Email inbox, gotten some groups summarised already and cross communication sent out.

Congrats on the launch!

Loving the new version! Consumer AI is only really effective when it's truly personal, and Pally definitely nails that 👌

I've been trying Pally & it's a genuinely polished experience. The onboarding is smooth, the UI is very clean, and it solves a real problem without unnecessary complexity. It feels like one of those products that you can start using immediately instead of spending hours figuring it out.

Congrats to the team @hazhubble & @wylans on the launch, looking forward to seeing where you take it! 🚀

'nobody wants relationships in another tool' is the honest pivot. but i don't have one tone, i text my mum unlike a client. does pally learn tone per person?

Texting an AI assistant over iMessage feels weirdly natural, like texting a friend who actually gets things done. Pally picked up my calendar and inbox without me having to fiddle with setup, which was a nice surprise.

claude.md ingestion so you can drop it entirely is the right call, most of these tools just bolt on another layer on top of it instead of replacing it.

Super exciting - congrats on the launch! A self-updating wiki for coding agents is a genuinely clever idea.

Congrats on the launch, Kushal. We're a small team and do a bunch of stuff with coding agents daily and the thing you're describing is a real problem. Anchoring facts to committed code so they expire on their own is the part I like, because it's the only version of this that doesn't turn into another doc nobody updates. Adding Greplica to my list!

Excited for the launch! Can this help and scale at enterprise level too?

Can I see exactly what it has learned and remembers, and edit or expand it if necessary? For example, in the form of concise documentation or something similar.

Anchoring facts to committed code is the part that convinced me — it gives stale knowledge a natural expiry date, which hand-written docs never have. I maintain around ten repos on my own and my claude.md files rot faster than the code does. What I couldn't tell from the page: how does Greplica sit next to an existing claude.md or agents.md — does it replace them, feed them, or end up competing for the same context window?

@kushal_patil Congrats on the launch! with multiple clones and forks feeding the same wiki, how do you handle two forks that genuinely diverge in approach?

Kushal, the quiet way teams forget why their own code works the way it does has bitten me more times than I can count, and this speaks straight to that.

Really nice concept for Greplica: Self updating wiki for coding agents. Rooting for you today—congrats! 🚀

Very interesting. Definitely needed for teams shipping real products at inference speed.

Congratulations! Turning all those little codebase lessons into something reusable is a great idea. That tribal knowledge usually gets lost way too easily.

Great stuff, this is a much more realistic way to make context available to new developers and coding agents.

Anchoring facts to committed code handles the "what does this do" half. The half I'd worry about is "we tried X and it didn't work", because that isn't anchored in anything and it's the note that costs the most when it's wrong. One bad conclusion from one bad session becomes ground truth for every agent run afterwards, and it reads exactly as confident as a true one. What retires a fact, other than the code it points at moving?

something i always wanted my coding agents to do

Can this also be used by product managers for data analytics? Will it be more efficient?

Code graphs usually capture what calls what. Pulling decisions, failed approaches, constraints, and workflows from actual coding sessions feels much more useful. Kudos to the team!

Congratulations on the launch 🎉🔥

Does this work only with Claude Code, or can a team share the same memory across Claude Code, Cursor, Codex, and Copilot?

On April 17, 2026, we launched the first version of Expert Chase as a web app.

Since then, we've been listening closely to every piece of feedback, learning how people actually wanted to use Expert Chase in their everyday lives.

Today, we're excited to ship our biggest update yet.

Expert Chase is now available on Web, iOS, and Android, delivering a completely new native experience with support for both the Apple and Google ecosystems.

We've also introduced our new native integrations, including Apple Health, and Google Health Connect.

But this is only the beginning.

E.Y.E. is designed to become your personal intelligence for everyday life, learning from your world to help you stay organized, focused, and one step ahead.

Anything you can do in the Expert Chase app, your AI, E.Y.E., can do for you.

Thank you to everyone who supported us from our first web launch. Your feedback has shaped every step of this journey, and we're excited to keep building with you.

We'd love to hear your thoughts, answer your questions, and learn what you'd like to see next.

■
— Yair Cohen

Regarding tracking money and access to the bank app – is it safe?

■ App Store(iOS): https://apps.apple.com/us/app/expert-chase/id6766313000
■ Google Play(Android): https://play.google.com/store/apps/details?id=com.expertchase.app
■ Website:

https://expertchase.com
■ Web App:

https://app.expertchase.com

Download now:
iOS: https://apps.apple.com/us/app/expert-chase/id6766313000
Android: https://play.google.com/store/apps/details?id=com.expertchase.app

Which integration do users end up relying on the most after their first week?

The desktop version still stays, right? When I visited the website it had only two options iOS or Android.

@eye Congratulations. And happy product launch.

nice launch! and props for a backronym that actually means something, most of them are painful.

going after all of it at once is a big swing though. which use case are people sticking with first? with these everything-apps there's usually one habit that hooks people and the rest comes later.

Nice launch! How do you plan on scaling this up?

Love the vision of making AI genuinely useful for everyday life. Congrats on bringing it to iOS & Android! 🚀🚀

Congrats on the launch! Expanding from web to native apps is a huge milestone. What's next on your roadmap?

Congrats Yair. The Apple and Google integrations make this feel much more useful in everyday life.

I remember you launched on Vercel Day maybe for the first time. Congrats on the re-launch.

Hey everyone 👋

I’m excited to share something I’ve built called Focus Room!

I built Focus Room because YouTube has some of the best educational content online, but recommendations, Shorts, comments, and unrelated videos can make focused learning difficult.

Key features:

• Open a video or playlist in a Coursera-inspired learning interface.

• Follow structured courses created from educational YouTube playlists.

• Generate summaries, take timestamped notes, and manage learning tasks.

• Track your course progress and bookmark courses for later.

• Add focusroom.club/ before a supported YouTube URL to open it directly in Focus Room.
  

This will be very needed for the sciences. I like this educational idea :)

Congrats on the launch!
Absolutely great idea to turn youtube into structured learning material.

P.S:
I have also built a quick livedemo for you
https://app.livedemo.ai/livedemos/6a6af926cc86a442a8fd9cd8

Hi, i like your idea and am very interested to try this app. But i got an error:"We couldn’t load playlist results.

Quota exceeded for quota metric 'Search Queries' and limit 'Search Queries per day' of service 'youtube.googleapis.com' for consumer 'project_number:492833098910'."

Product suggestion: The page layout needs to be optimized. The video is the core element. However, in the default opened page, the video area occupies too little space. If users switch to full screen, they will not be able to use the provided tool to record the content.

the quota exhaustion someone hit above is the thing i would design around first, not a launch day bug. anything sitting on the youtube data api inherits a hard daily ceiling, and search is the expensive call, so your cost scales with usage while the product is free.

caching resolved playlists and topic maps per video id rather than per user would buy a lot of headroom. two people working through the same calculus playlist should only cost you once

Congratulations on the launch! The most useful business advice I found was on YouTube and in books, and the problem with the YouTube part was that none of it stayed anywhere. I would watch something useful and then rewatch half of it later to find the one section I needed, so timestamped notes solve most of that for me. Can those notes be exported or do they live inside Focus Room?

I was genuinely excited to try this out, but I ran into a couple of issues almost immediately. The first one was during sign-up. When I clicked the email confirmation link, it redirected me to:

http://localhost:3000/?error=access_denied&error_code=otp_expired&error_description=Email+link+is+invalid+or+has+expired

At first I assumed the account wasn't created, but I was actually able to log in afterward, which made the flow pretty confusing. Then, when I tried searching for something, I got this error instead:

We couldn’t load playlist results.

Quota exceeded for quota metric 'Search Queries' and limit 'Search Queries per day' of service 'youtube.googleapis.com'.

It looks like the app has already exhausted its YouTube API quota for the day.

I still like the idea behind the product and was looking forward to trying it, but so far I couldn't

I could see this being useful for university students too. Have you had many students using it for lectures and study playlists?

the paste-based insertion is the part i'd poke at. macOS deliberately blocks the accessibility APIs from reading or writing into secure text fields - password fields, a password manager's own vault unlock, sudo prompts in terminal - specifically so assistive apps can't fill them. does yap detect it's in a secure field and just decline to paste, or does the hotkey still fire and it dictates into empty air / fails silently there? asking because "works anywhere you can type" and "macOS explicitly carves out a class of fields no assistive app can touch" are going to run into each other somewhere.

fire

Congrats on the launch, this looks awesome! On-device, open-source dictation for Mac is exactly what I've been wanting.

Great App. @Wispr Flow killer :-) . I tried and discontinued wisper because I thought talk to text should be part of the OS as a free feature. Looks like you guys solved that problem. One issue is verbal stumbles. If you can fix that, it will be a super cool.

What languages are supported?

that's a great product! + after everything else being bloated, a few MB of memory and no network calls feel like such a relief.

Paste is the part I'd want to hear about. If you're putting the transcript on the clipboard to insert it, you're stomping whatever I had copied, and that's the bug that gets a dictation app deleted on day two. Restoring the previous pasteboard afterwards mostly works but races with anything else watching it. Separately, your description says it idles near 60 MB and your comment says a few MB, worth picking one.

Hey PH, we built agentOS because every time we shipped agents to production we ended up rebuilding the same stack: code execution, file storage, orchestration, permissions, state that survives restarts. That usually means a sandbox provider, object storage, a workflow engine, and a lot of glue code.

agentOS packages this as a single library. Each agent gets its own lightweight virtual operating system running inside your existing backend process:

Execution: Node.js on native V8 (full JIT, not JS compiled to WASM), Python, Bash, and subprocesses with Linux-like semantics. The agent writes one program instead of chaining tool calls.

Filesystem: persistent POSIX filesystem. Mount S3, Google Drive, or host directories at normal paths and use regular files and shell tools.

Orchestration: durable workflows, crons, shared sessions, human-in-the-loop approvals, agent-to-agent. Written as normal application code, checkpointed automatically.
Control: expose typed backend functions to agents without handing over credentials, review permission requests in your own UI, cap resources per VM.

It runs on WebAssembly so it's small: ~4.8ms cold starts and ~22MB per agent instead of a dedicated VM per agent. If a workload needs something more, you can mount a sandbox.

Works with Claude Code, Codex, OpenCode, and Pi, or bring your own agent. Apache 2.0, one npm install, deploys wherever your backend already runs.

Happy to answer questions about the architecture.

Really exciting release! I'm a big fan of the Rivet team and everything y'all put out. Looking forward to trying this out because we keep scaling the number of background agents and tasks but performance and compute is a huge pain point!

PS this team is ahead of the curve and open source. I even had a contribution merged in a day or two!

Nathan and team are creating something beautiful. Been keeping my eyes on AgentOS and super cool it works with Eve and/or Flue

agentOS is 8 months the making building what we believe to be the future of secure agent isolation in a lightweight library.

We're excited to finally share this with you all!

The number I would want next to 254x cheaper is what I give up to get it.

WebAssembly buys you the cold start and the cost, but it is a different syscall surface from a real container. Most agent work is not clever, it is shelling out: git, package managers, native modules, something that wants to fork a process or open a socket the way it always has. The interesting question is not whether the sandbox is fast. It is which of those quietly fails, and whether it fails loudly or just behaves oddly.

A compatibility page would do more for adoption than another benchmark. Here is what runs, here is what does not, here is what runs but slower. People building agent harnesses have been burned by "mostly compatible" before, and the ones who have will look for that page first.

Genuine question: when something is unsupported, does the agent get a clean error it can reason about and route around, or does it get something confusing? That difference decides whether an agent can recover on its own or just loops.

Sounds interesting. How easy is it to install? Can I do it myself, or do I need an administrator?

254x is doing a lot of work in that tagline and the number isn't anywhere in your comment, so say against what. Firecracker per agent-hour at list price is a different claim to a per-request one. Some of that saving is also the isolation boundary you removed, since in-process WASM means a bug in the host bridge lands inside my backend instead of in a VM I was already paying to throw away. I'd lead with 4.8ms and 22MB, those are checkable.

Hey Product Hunt 👋

We're the team behind OpenCV (the open-source computer-vision library). For the past two years, we’ve watched AI video get sharper but stay short — Seedance 2.0 tops out around 15s, most models do <10s, and the moment a clip runs long the face drifts and the illusion breaks. That's useless if you actually want to say something — a product walkthrough, a training module, a spokesperson.
So we built CraftStory to solve length and identity together. Give it one photo and a script (or your own audio), and it generates coherent, minutes-long talking videos — same face frame-to-frame, accurate lip-sync, natural gestures, in 30+ languages. You can also:

• build a custom avatar of yourself from a ~15s video,

• pick from 100+ ready-made AI actors,

• or hand the whole thing to a Video Agent that writes the script, adds b-roll/subtitles, and casts an avatar — all edited by chat.

There's a free plan (no card) so you can try it on your own photo right now: https://craftstory.com/
We'd genuinely love your feedback — especially: if you could generate a minutes-long video with consistent characters, what would you do? And where does it still fall short? We're in the comments all day.🙏

1) What is the difference between Model 2.0 and AI Actors? Am I correct in understanding that Model 2.0 is used to generate any type of video scene? Could you explain in more detail what AI Actors can do and when they should be used instead of Model 2.0?
2) Is it possible to purchase additional Model 2.0 credits if I need to generate more videos than my monthly credit allowance?

Looks good, and the opportunity to test it for free is nice! Not exactly what I was looking for (not UGC- or avatar-type, but more creative videos for promoting products/services), but should be a good product for its purposes.

Awesome to see this launch happening and coming from the team with some of the OGs of the computer vision industry makes it even more exciting.

Good luck guys, love the product and the tech behind it, exciting to see where you take this forward! Kudos to Victor and Alex!

Congrats on the launch!
AI avatar videos are getting genuine traffic, my parents told me last week about videos they watched and they didn't realize they were AI made.
I guess good content is good content no matter how you package it!

P.S:
I have made a livedemo for your website
https://app.livedemo.ai/livedemos/6a6afb2ecc86a442a8fd9d56

Licensing the actor footage through revenue share is the part buyers will care about, and it's buried at the bottom of the description. L&D is exactly where legal asks who is in the frame and what happens if that person objects in two years, and a licensed model is a very different answer to a scraped one. I'd put that line above the 4.5 cents.

Awesome, the genai space is heating up even more :) multi-minute consistency is an impressive feat.

finally something that tells me my Claude session is about to hit the context ceiling before it silently stalls. the little cat is silly but the live meter actually saved me from a wasted run this morning.

the whole premise is "you switched tabs and forgot" - but the cat itself lives in a corner of a screen you're, by definition, no longer looking at once you've switched away or stepped out for a meeting. does an alarmed session ever escalate past the widget itself, a system notification, a sound, something that reaches you if you're not glancing at that corner, or is the ceiling of the alert exactly the failure mode the product exists to catch, just wearing a cuter shape?

building it as a Tauri widget instead of an Electron dashboard is the right call for something meant to live in the corner all day. what's idle memory actually look like with 3-4 sessions being watched at once?

The moods are the risky part. A tool approval that's blocking right now and a context window that fills in ten minutes are different urgencies, and one alarmed cat flattens them into the same glance. With four sessions running I'd want it to only wake up for the blocking one, otherwise it's alarmed most of the day and I stop seeing it.

The per-session context meter is the right primitive. Usage dashboards tell you what you spent; this tells you what's about to break, and those are genuinely different problems.

One thing I'd want once several are running: sessions aren't equally valuable. I route deliberately — cheaper models doing gathering and research, expensive ones doing judgement and synthesis. A gather session hitting 90% is fine, it's disposable, I'll rerun it. The synthesis session hitting 90% is the expensive one, because its context is the work product, and an auto-compact there quietly loses reasoning I can't get back.

A flat meter treats those identically, so the cat gets alarmed about the wrong one. Some way to mark a session as the one that matters — pin it, priority tag, whatever — would make the alarm mean something. It'd also make subscriptionbox's 75% heads-up land harder, since the early warning is really only worth having on sessions where planning a handoff is worth the interruption.

Leopold's question is the other one I'd want answered — whether pane-to-session mapping survives a restart is the difference between a widget I leave open and one I re-set-up every morning.

The mood states are a good call, by the way. Glanceable beats precise for anything living in peripheral vision.

The context-window meter per session is exactly what I kept wishing existed. I always notice too late that an agent quietly compacted. Does the widget remember your session layout between restarts, like which terminal panes map to which agent, or does it rediscover them fresh every time?

This hits close to home. I run long Claude Code sessions for a data pipeline and losing track of a session mid-context-window is exactly the failure mode you're describing. The mood states (sleeping/running/alarmed) are a nice touch, way more glanceable than a numeric meter. Following to see how the Windows/Linux support shakes out.

A tiny cat watching my coding agents is genuinely delightful — and so useful. Congrats on the launch! 🚀

I feel like coding agents are becoming something you supervise instead of actively use. I'm curious if you've thought about adding simple stats, like which agent spends the most time waiting for approvals or burns through context the fastest.

love the cat mascot idea, super cozy vibe honestly. one thing i'd want though is a little sound or subtle notification when a session is about to hit the limit so i can wrap things up before it gets cut off mid task. right now i think it just nudges you right when it's already too late, like a few seconds before the cutoff. maybe a heads up at like 75% would be way more useful so you can actually plan the handoff

Hey Product Hunt ,

Generative AI made fashion content cheap. It didn't make it usable.

Generative AI enabled you to create 1 image at a time. It didn't build actual bulk workflows for you.

If you've ever put an actual garment through an image model, you know the failure mode. The drape is wrong. The fit reads two sizes off. The color is off, the background is changing in every image, the model is not consistent and phew!! you need to do this image by image to get it right.

My co-founder Kirti ran over 200 physical photoshoots as CEO of Okhai. Every new style meant the same grind — ship samples, cast models, book a studio, wait on retouching, reshoot because the fit read wrong on camera. About $4,000 and several weeks per style.

Switching to AI seems like a no brainer, but tools are failing to provide consistency or the speed AI is supposed to bring to our workflows.

So we built Caimera: an AI visual production platform that takes a garment from sketch to sale.

• Design — sketch to image with real texture, print and textile accuracy. Tech packs in minutes. Trend-ready designs from your own brand DNA. Batch produced.

• Ecommerce — flat-lay or sketch to a full on-model catalog, consistent backdrops, lighting and models. Batch Produced. Caimera also has a tool for creating ghost mannequin images, recoloring with texture intact, and batch resizing and background change.

• Marketing — create editorial imagery with the largest library of AI fashion models, 20,000+ creative templates for editorial images and product videos. Caimera also supports 14K upscale for billboards.

Where it's landed so far: 99.3% lower cost per image, 10× faster to publish, ~50% lift in CTR. Teams at H&M, Steve Madden, Superdry, Dolce Vita, Kurt Geiger, Ioni Swim and Floafers are running their AI visual production pipeline on it.

What I'd genuinely love feedback on: where the workflow breaks apart. Bring us your hardest bulk workflow, you want to generate 100 sketches, then create tech packs and lifestyle images for all the styles to launch, but the workflow is shaking — tell us where we fall short. That's the only benchmark that matters to us.

Kirti and I are here all day.

— Prateek, co-founder & CTO

Tried a few tools in this space, Caimera’s Editorial templates is what made it stick for me.

• Caimera‘s Upscale feature is the cherry on top, with up to 14x upscaling

• Blown away by the speed, quality and ease makes high-end photos and videos. It saves so much time and money

• The built-in video editor makes stitching together sleek product clips super fast. Really impressed with Caimera. 

This tool has the best best model selection. They even have a custom model making option which I tried my hand at, sooo easy to use. But the selection itself has such Authentic realistic variety of faces, body types; all with amazing skin texture, makes my work really stand out and look so edgy!

Why was I making life so hard before? 😭 Like literallyy caimera saved me HOURS of work. Lowkey obsessed with how smooth this workflow is.

This looks fantastic for asset creation. However, most e-commerce platforms have specific dimensions, product placement and visual guardrails that can sometimes reject assets from high traffic placements (mastheads, etc.). The consequence of this is a longer lead time, missed branding opportunities and inefficient utilization of the fixed buy period. It would be fantastic if that was programmed by default for leading retailers in key markets. That way, it would solve 2 key challenges - asset creation and asset approval (best practices by the platform) in the same workflow.

The failure you describe is the expensive kind, because a wrong drape still looks like a good photo. Cost per image and CTR both look fine when the fit reads two sizes off, and you find out at the returns desk six weeks later. If H&M or Steve Madden will let you publish return rate on AI-shot styles against photographed ones, that's the number that ends the argument, not 99.3%. The rest is a procurement conversation.

We sell into indie consumer brands, so this is aimed squarely at people I talk to weekly. Two things from that side of the table.

The buying trigger is almost never "our photos could be better". It is a launch date with no shoot booked, or a marketplace listing rejected for image spec. Founders do not shop for a visual production platform. They shop for the thing standing between them and going live on Thursday. Whatever the equivalent of that is for you, that is the page they should land on.

The objection you will hit is not quality, it is whether they can use the output. Amazon, Shopify and the big marketplaces each have their own image rules, and a beautiful asset that fails a listing check costs more than no asset. If you already handle spec compliance per channel, say it loudly, because it is worth more to a small brand than another style preset.

The synthetic person label on that first image is a good instinct, by the way. Getting ahead of disclosure before anyone forces you to is going to age well.

open-source motion tool you can actually fork is the rare bet, not another closed box. does the ai drop editable keyframes, or a result you can't reopen?

Determinism as the headline is the right call, that's what people actually get burned by. The thing that'll bite you today is Windows only, most motion people are on Apple silicon and 'coming soon' on a download page reads as 'not yet'. Also you're calling it Motion Editor here and Premation everywhere else, worth picking one before the day fills up.

Really like how Laxis goes beyond simple transcription. The combination of AI note-taking and voice dictation feels like a real productivity boost, especially for people who spend hours in meetings or constantly switching between writing and speaking.

to be honest, love the part of no bot joining the call the most. it was always very weird for me with other note takers. good product!

Me believung the combination of transcription translation and dictation makes this more than a simple meeting tool. The feature set covers several dailt tasks. Case studies showing measurable time savings would make the overall value even easier to understand.

The auto-share default flagged in your reviews is the thing I would fix before anything else on the roadmap.

A summary that lands in every attendee's inbox changes what gets said in the room. People already talk differently once a recorder is running. If they also know their words get summarised and distributed automatically, you get a more careful and less useful meeting, and you never see the cost, because the thing you lost was never captured in the first place.

On sales calls it is sharper than that. My internal read on a call is not something I want auto-delivered to the prospect with a confident summary voice attached to it.

Separate thought on the positioning. Note-takers cannot dictate, dictation apps cannot run meetings, so you do both. I spent a year selling a catalogue and the lesson that cost me most was that buyers do not buy the union of two things, they buy the one they already came for.

Which of the two is the reason people actually stay? From the reviews it reads like the SDR side, which would be a third answer again.

The workflow feels intuitive. I can easily see this becoming part of my daily meeting routine.

honestly love the hover preview idea, that alone makes it worth a try. one thing though — would be great if the widgets could be resized or reordered, because depending on my screen setup the music player ends up way bigger than the weather and it looks unbalanced. small QoL thing but makes it feel more like my setup

Excited to see docktor live. Your Dock's wasted side space, now full of widgets is a compelling value proposition—best of luck today! 🚀

free, notarized, no account, solo dev reading everything. you make it easy to root for you!

Nice use of space nobody thinks about. What I'd want to know is what happens when the Dock grows, because it recentres every time an app opens and the side space shrinks with it, so the widgets either reflow or get clipped mid-glance. Same question when the Dock jumps to a second display.

the Text Area Portal idea is genuinely clever, basically turning any input field into something you can edit in your actual editor instead of fighting with cursor placement. feels like a really thoughtful gap to fill.

Predictable rather than arbitrary hints is the whole thing, and I am glad it is in the first line instead of buried in a feature list.

Arbitrary hints mean you read the screen every single time. Predictable ones mean the sequence moves into your fingers after about a week, and that is the difference between a tool people demo and a tool people keep. Most of this category optimises the first press. You have optimised the hundredth.

Text Area Portal is the one I did not know I wanted. I write long text into browser textareas constantly and it is miserable, and every workaround I have tried means composing elsewhere and pasting back, which loses the field state and half the time the formatting.

Question rather than a critique: does the hinting read the accessibility tree, or something else? Apps with missing or lazy accessibility labels are where everything in this category quietly falls over, and I would rather know upfront which ones you already know are rough than discover it myself on a Tuesday.

This would be quite useful for humans as well as computer use agents. Computer use relies on taking a screenshot and sending it to an LLM. In this case, an agent does not have to rely on mouse movement to take an action on the screen, as it can simply use the keybindings shown in the screenshot.

to answer the question: for me it's remembering the small stuff between the big moments, someone mentioned their kid was starting a new school or they were nervous about a specific presentation, and that's exactly the detail worth bringing up three months later, but it's also the first thing to get lost because it never felt "important" enough to write down at the time. every relationship tool lives or dies on whether capturing that is fast enough to actually happen in the moment. how are you getting people to log that kind of detail, quick text entry after a call, importing from calendar/email, voice notes? if logging takes more than a few seconds people just won't do it consistently, no matter how good the recommendations on the other end are.

this is a great product! and to answer your question, I'd say follow up. staying in touch when there's no ask.

The line I keep turning over is "without turning relationships into transactions", because the mechanism underneath it is ranking people by what they can do for you.

Recommending who matters and why they matter is a scoring function over your friends. That is not a criticism of building it. It is a question of whether the product can say plainly what it does without the user flinching.

Worth noting Pally launched on this same board today, and they rebuilt away from exactly this. Their stated reason was that nobody wants to manage relationships in another tool, they want the follow-ups to just happen. Two products, one day, opposite conclusions from the same problem. That is interesting rather than damning.

Real question, not a gotcha: what happens to someone who scores low? Do they get quietly deprioritised, or does the ranking stay behind the scenes and never get shown to me at all?

This is super exciting - congrats on the launch! A relationship system that focuses on context over transactions is a refreshing take.

Congrats on launching Virre! A private relationship system for your career and network. is an exciting direction, and I’m looking forward to seeing where it goes. 🚀

the framing around "who matters, why they matter, and what to do next" feels really thoughtfully sequenced. so many tools jump straight to action prompts and forget the context layer that actually makes you trust the suggestion.

no bot joining the call is the detail that got me, most notetakers force that visible presence. if it's just reading system audio instead of joining as a participant, how do you handle speaker attribution on a 4 or 5 person call?