PH热榜 | 2026-08-27

Hey Product Hunt! I'm Marcus, Co-founder of Skydive 👋

The problem

Everyone has access to the same AI models now, but you still have to prompt AI tools to get good results. I've seen one too many terminals built for managing Claude Code instances.

Today's tools usually fall into two camps:

💬 Chatbots answer questions and generate content, but you still have to take action yourself. They forget what you taught them the last time you worked together.

⚡ Workflow builders automate repetitive processes, but you have to design every step, maintain the logic, and update it whenever something changes.

Skydive is designed to actually help you grow your business by autonomously taking action throughout your company.

Meet Skydive

Skydive lets you hire AI agents that take on real responsibilities across your company.

Describe the job, and your agent gets to work. Your agent has it's own codebase and computer in the cloud, but you never have to think about it. It just works. If you want it to use your skills, you just tell it to. If you want it to learn to write like you, it can do that too. It also self-improves over time by learning from previous conversations.

🖥️ Every agent has its own computer

Agents use websites, apps, and files just like a person would. They click, type, log in, create documents, and complete work from start to finish.

💬 Agents work where you work

You can talk to agents in Slack, email, iMessage, on the web, or in your terminal. They go wherever you go and feel right at home working alongside you. They even connect to your desktop and can use the same programs you use.

🤝 Built for teams

Agents are experts in their own area. They collaborate with each other, share context, and hand work off automatically so bigger projects actually get finished. Every agent can be shared with other members of your team so multiple people can coordinate work.

🌙 Automation that never sleeps

Turn recurring work into routines. Your agents monitor, take action, and keep work moving whether your laptop is open, closed, or you're halfway around the world.

🧠 Agents that improve over time

Correct an agent once, and it remembers. Your preferences, feedback, and company knowledge carry forward into future work automatically. When you're done working, your agents are dreaming and ingesting the lessons from the previous day's work.

Who is it for?

Founders, operators, and fast-moving teams that need to do more without adding headcount. If your work spans multiple tools, teammates, and recurring processes, Skydive was built for you.

We'd love your feedback ❤️

We're just getting started, and we'd love to hear what you think.

If there's one responsibility you'd hand off to an AI agent, tell us in the comments. We'll be around all day answering questions and shipping improvements.

To celebrate our Product Hunt launch, we're giving everyone an additional $5 in credits with the code PRODHUNT5 - if you need more, shoot my cofounder an email zaria[at]anything.com :)

If you're launching on product hunt soon, you can use this agent we've built to organize your launch. We used it ourselves, so we'll see how this goes, lol

https://www.skydive.com/templates/product-hunt

Thanks for checking out Skydive! 🚀

“Cloud agent coworkers " is a great way to frame this. If this agents actually get better through repeated work, that could be a huge productivity boost.

Congrats. Are we able to control permissions to agents? Example, access to production tools and company data?

While my customer facing support agents will always be my favorite part of my workday, its the niche agents like ChargeKnight (Stripe Disputes Manager) and Alfie (Affiliate Fraud Prevention and Payments Manager) that win for me on those areas of the business I care about but don't work in daily. They build expertise off best practices (white papers on Stripe Disputes) from external sources and adopt them into our company - Chargeknight even pays for himself, he returns more cash to us than he costs to operate by a factor of 2-4x

Alfie identified spoofing in our affiliate program with pending payouts of over 18k. The vigilance of Skydive agents helping me around the clock is just unbeatable.

Is Telegram support on the roadmap?

pro tip: make an agent that looks similar to you and train them to talk and work like you. then tell your coworkers to go through the agent before escalating to you. congratulations you just made your personal assistant.

huge congrats to dhruv and the whole skydive team on the launch. been following this since i signed up - agents that actually do the work across your tools instead of just chatting about it is exactly where this needs to go. pumped to dig in. rooting for you guys

I like that the focus is on outcomes rather than prompt writing. Not everyone wants to spend time figuring out the perfect instructions.

How does the tool decide when to rely on an agent and when to ask human input? Or that totally depends on the task?

Skydive looks incredible, congrats on the launch!!

Congrats on the launch. Nice touch running today on the same launch agent you're handing to everyone else.

Running cloud agents that actually hook into your existing tools without a hassle fixes a huge orchestration pain point. Congrats on shipping!

This feels especially useful for repetitive work that usually gets stuck between different tools. How much control do teams have over what an agent can access and change?

Love how easy it is to create helpful working agents with skydive. Most productive couple days I’ve had in a minute!

So pumped for this. I've been contemplating looking for a co-founder. This is my sign to build Skydive agents!

The fact that these agents work across existing tools is what interests me most. It feels more practical than adding another separate workspace.

doing dev work with a skydive agent is underrated. I literally take my work from the web, to the CLI, to iMessage on the train. the grind don't stop even after you leave your computer.

craziest part is still the fact that slack is now a multiplayer IDE, CRM, support triage workspace, and everything in between for us. I don't ever want to go back

Hey everyone! I'm Rain, one of the designers on the team here. I've been designing Skydive for the past few months and am so happy to see it live!

I created a design agent named Echo who has greatly boosted my productivity and hope everyone gets that magical experience too. Feel free to let me know any product and design feedback you might have :)

(and share your agent card with me once you create one - I'd love to see!)

Congrats on your launch but I need to understand what it's all about and how it can help me

We're really out here hiring cloud coworkers now hahaha. Love the idea of just describing what needs to get done and letting the squad handle the rest. Cool stuff guys, congrats on the launch! @zariazinn @dylan34 @genevieve_tankosich

Congrats on the launch guys! Seems like a pretty solid productivity tool.

The multi-agent collaboration concept is probably what I’d explore first. Having specialized agents share context and hand work off to each other could be very powerful for larger projects. Curious to see how well this works with real-world, messy workflows.

What caught my attention is the persistent memory across different platforms. Being able to start a task in Slack and continue it elsewhere without having to explain everything again could make AI agents genuinely useful for day-to-day operations.

The idea of giving an AI agent its own cloud computer is really interesting. Being able to let it actually interact with websites, files, and different tools feels much closer to having a real digital teammate than just another chatbot.

Hey Product Hunt 👋 I'm Eric, one of the makers of Enter Pro.

Today, we're bringing Enter Pro to Product Hunt for the first time.

As AI models grow more capable, the barrier to coding keeps falling. But building what runs a business takes more than code.

Enter Pro is an AI-native platform that turns idea into working apps. As the builder layer of Converge AI, teams can plan, build, preview, launch, and scale apps, websites, portals, internal tools, workflows, and AI agents in one continuous workspace—with databases, authentication, storage, hosting, payments, analytics, and localization built in.

Enter started at the beginning of the year as a small experiment built around one question:

Can Enter build Enter?

We became our own most demanding customer. Our blog system(CMS) was the first product built entirely inside Enter. When GitHub import launched, we brought our main repository into the platform and began building Enter with Enter. Product, operations, and design could ship smaller improvements directly, while engineering stayed focused on the deeper systems behind them. Over time, more of Enter was built on the platform itself—from the Enter Forum and our internal admin system to landing pages created and shipped by non-technical teammates. The answer was yes: Enter could build more than prototypes. It could power the real systems behind our business.

Along the way, we noticed that our teams were still repeating the same work every week—answering support questions, researching accounts, summarizing feedback, and preparing updates. We began turning those workflows into agents, which led to one of Enter Pro’s newest capabilities: Agent Builder.

Describe what your agent should do, test it in a live preview, and publish it as a shareable link, a web agent, inside your product, in Slack or Lark, or through an SDK or embed. Enter handles the models, knowledge, tools, deployment, versioning, and infrastructure behind it. The goal isn’t another chatbot or one-off demo. It’s to turn recurring work into agents that can be deployed, maintained, and improved over time.

What excites us most is the shift from AI as a chat window to AI as an operating layer. And because apps and agents are built on the same platform, AI can become part of how a product works from day one—not something added afterward. Our belief is simple: people should spend more time on judgment, creativity, and deciding what to build—and less time repeating the work required to make it happen.

We still have a long roadmap ahead, and we’d love to hear what you want to build with Enter.

As a small launch gift, the first 300 new Product Hunt users who sign up and try Agent Builder will receive 400 bonus Credits. First come, first served—claim them within seven days 🎉

👉 https://enter.converge.ai/s/ph1

love the can enter build enter? story! how has using your own platform changed the way your team develops new features?

Congrats on the launch, Eric!

What makes Enter Pro different to those big players like @Lovable and @Base44 ?

How does collaboration work when several people are working on the same project? Is there version history or branching?

Congrats team! what types of AI agents have you found to be the most valuable in your own daily operations so far?

Hey Product Hunt 👋 I’m part of the team behind Enter Pro.

I worked mainly on shaping the Agent Builder and discussing the backend architecture—especially how agents connect to knowledge and tools, get deployed and versioned, and evolve beyond one-off chatbot demos into reliable workflows.

The part I’m most excited about is having apps and agents built in the same environment, so AI can become part of the product itself rather than something bolted on afterward.

Huge credit to the engineering team for bringing it all to life. What’s the first recurring workflow you’d want to turn into an agent?

Congrats! @eric_zhang25 . I am Leo founder of flaq.ai. It's very interesting to build apps with Enter Pro. Hope that more and more builders can find Enter Pro to help their building jobs.

Having apps and agents in the same product makes sense. AI is more useful when it is part of the workflow, not a separate chat box bolted on later.

Congratulations on the launch! Is it like Lovable, but with ready-made modules?

this looks cool! is this a web app or a desktop app? or both?

Congrats on the launch! A lot of AI builders are great at getting you to a prototype, but the harder part is everything that comes after—auth, databases, payments, deployment, analytics, and actually maintaining the product. Bringing all of that together with an agent builder in one workspace feels genuinely useful. Curious to see how Enter Pro handles more complex apps as they scale.

Interesting product. Is this for product managers?

Congrats on the launch. Building the platform inside the platform was the hard way to prove it.

Congrats on the launch! Being able to build and scale full native apps using AI is huge for anyone trying to move fast. The workflow looks super clean and straightforward.

Does it have the ability to connect to multiple connectors? I mean, integrating with other tools?

What base models does Enter Pro use?

I have seen tools that are good at just one thing. Example, yesterday's top product was a tool that helps build only iOS apps really well. Whereas Enter Pro seems to build everything and anything. What do you think is the baseline difference between horizontal versus a vertical builder tools?

I like the idea of custom AI agents being part of the same app-building environment. That could make internal tools much easier to create.

Can teams switch between different AI models depending on the task, or is the model choice handled automatically?

Having localization built in from the start is a thoughtful touch for teams planning to serve multiple markets.

I’d be curious to see how it handles an existing app rather than starting from a blank project.

Congrats on the launch, looks really promising!

What happens when the AI makes a change that breaks an existing feature? Is there an easy way to inspect and fix it?

I really like the continuous workspace concept. how seamlessly can you move from planning an idea to deploying the finished product?

Kosta here, co-founder of Lenz. Many businesses ship AI-generated content to their customers. Some of those use cases could benefit from factual verification of that AI output. That's why we built Lenz, packaged it as an API/SDK, and made it available across multiple platforms (n8n, Zapier, MCP, CLI), so people can easily integrate it into their workflows. Lenz verdicts come with a full audit trail - sources, citations, reasoning, confidence. 

How Lenz is different than just asking a model:

(1) separate evidence gathering step (with source ratings) that doesn't rely on the model's memory or retrieval capabilities

(2) multi-vendor, multi-model approach to address single-model biases

(3) multi-round adversarial debate to crystallize the strongest for/against arguments

(4) multi-model jury reviewing the evidence and the debates across multiple axes

Key API primitives: /extract - extracts the factual claims from a text; /assess - quick assessment of a claim; /verify - the full deep claim verification; /ask - follow-up post-verification questions.

We measured the level of disagreement between the individual frontier models: on 23% of real-world claims, they disagree significantly, which sets the floor of the error Lenz is built to address.

To try a Lenz verification via the UI: lenz.io/verify

More about the LLM disagreement research: lenz.io/research/llm-disagreement

To integrate Lenz: lenz.io/integrations

GitHub: github.com/lenzhq

Hope you find this useful. Let me know either way :)

Hi, Vicky here 👋 the non-tech co-founder of Lenz

AI has genuinely changed what I can do on my own (from ops to marketing to building alongside an engineering team without being one). I’m one of those people it has really opened doors for.

But there’s another side to this. More and more companies rely on AI to produce the actual content their customers see: reports, articles, recommendations, research, support replies, ... And "someone will check it before it goes out" stops working at scale.

That’s the part of Lenz that really resonates with me. For editorial and production teams, that means letting verified claims through, sending uncertain ones for human review, and keeping the sources behind every decision.

AI will let us produce vastly more. Verification has to scale with it.

Very excited to have Lenz out here today and would love to hear your use cases.

Hi, I'm Pavel, part of the Lenz team. I worked on the MCP server and the CLI amongst other things.

Quick note on why we believe this needs to exist: we gave the same 1,000 real claims from Lenz to 5 frontier models (Fable, GPT 5.6, Gemini 3.1 Pro, Sonar Deep Research, Grok 4.5), identical prompt, all with web search and thinking. All five agreed on only 37% of them, and on 23% the verdicts were 2+ steps apart on a 5-point scale. Surprisingly confidence was almost meaningless: 76% of all answers were self-rated at 9 or 10 out of 10. The figure below is that result. Each ring is a model, each spoke is a claim, grouped by agreement.

The two services I built:
- MCP server at lenz.io/mcp. To connect see lenz.io/integrations/mcp-server or just hand it to your agent.
- CLI: pipx install "lenz-io[cli]", then lenz verify "the Great Wall is visible from space"

Paper, data and prompts are open if you want to pick it apart: lenz.io/research/llm-disagreement/v1.1

Happy to answer anything. Let me know if you try the MCP.

Congrats on the Launch @kostaj This hits close to me, spent Years in Security Testing and the same problem shows up there: one Models blind spot becomes everyones blind spot if you don't cross check.

Congrats team..idea of having a jury review the evidence is quite interesting.

This feels especially important right now. We are living in a world where AI is becoming one of the first places people go to check what is true, yet your research shows that frontier models disagree on 63% of real-world fact checks.

That is such a powerful reminder that a confident AI answer is not the same thing as a verified answer.

Really love what you’re building with Lenz. The need for trustworthy verification is only going to grow as AI becomes more embedded in how we work, learn, and make decisions.

At this point, I’m just happy to have an AI that occasionally says, “Let me check that” instead of confidently making things up 😂

Congrats on the launch! Does your product provide the full chain of sources for manual verifiability? (not just direct links, but maybe logical facts if this -> then this -> hence this <source>)

Love the idea and positioning! Excited to see where it goes huge congrats on the launch

fact checking as an api is underrated 🔍 media or legal buying first?

Congrats on launch number two. Publishing the data and prompts for anyone to pick apart is a good look.

Congrats on a launch, team @kostaj @vicky_dodeva @david19782 @pavel_j - are you a family, wow?

I have a question regarding the product as well. Could Lenz verify outputs from Deep Research in ChatGPT?

Wow. That's super cool. I read pitch decks all day and a lot of what's in them is forward-looking "market will be $1T by 2030" kinda of claim. Does Lenz work for that too? They usually have a market data source as the basis for their claims.

Half the pages that rank for file format questions I work on are confidently wrong, and they're the most linked. How do your source ratings handle a topic where the crowd is wrong?

I’ve already discussed this direction with Bek, and we even hosted a deep-dive online meetup together about the development of voice AI.

That’s why it’s especially interesting to see how the idea behind Speko is now evolving into a real product. Choosing the right voice stack is becoming increasingly complex as new models and providers keep emerging.

I really like Speko’s approach: independent benchmarks, routing based on the specific use case and language, and no bias toward their own models.

Very interesting product and direction. Good luck to Bek and the Speko team!

Congratulations on the launch @abdikbek! All of us needed a voice AI OpenRouter for sure. Wondering how often you would refresh this list of APIs or include more providers.

What are the advantages compared to Google’s open-source libraries and other alternatives? Also, what about less commonly supported languages like Swedish or Danish?

P.S. I just saw a post here about a kids’ app. You should consider partnering with them - they have voice recognition for children and convert speech to text.

Love your product, guys. We are using it at SUN

I love when a product does all the things I hate doing. Congrats, Speko. Proud user here.

Super hyped!

Looks good, but would be nice if the site provided also actual voice samples of each model, e.g. just same text with type of voice prompt benchmarked accross different models.

This is super cool!

Congratulations Bek. I have a question since Speko decides the stack before the session starts to keep audio paths direct, how do you handle mid-call failovers? If a specific provider suffers a sudden latency spike or an outage mid-session, does the system stay locked into that stack, or is there a way to redirect without dropping the user's call

Most of the audio I deal with is iPhone voice memos, never the clean format the docs assume. What happens to the input before it reaches a model?

openrouter for voice is such a usefull idea 🎙️ which stt is winning rn?

Love the "show me the numbers and let me pick" approach haha. Makes choosing between all these models feel a lot less like guesswork. Nice work guys, congrats on the launch! @abdikbek

Garry Tan hunter? Nice!

Given open router, this makes perfect sense! What are the surprising factors that affect which voice model is best for the situation?

congrats on the launch Bek! This is cool

I’m happy to share this! 🚀 Gemini 3.5 Transcribe is a big step forward for natural voice interaction.

Instead of typing every thought, it works the way you actually talk:

• Handles self-corrections

• Removes filler words & formats clean text

• Understands natural intent & speaking style

• Works accurately in noisy environments

• Up to 3 speakers with timestamps

• 85+ languages, accents & dialects

On Gemini for macOS, you can use your voice to generate images, search, summarize, analyze files and more.

It also powers Rambler on Android, turning spoken thoughts into polished text with voice-based edits, corrections and style changes.

🚀 Rolling out now on Gemini for macOS and Rambler on Android. Developers can build with it in Google AI Studio and Google Antigravity.

Congrats on the launch. Good to see a model that cleans up the filler rather than transcribing every word faithfully.

We’ve been building Traccia because we kept seeing the same gap with AI agents: once an agent can call tools, make decisions and take actions, a traditional trace can tell you what happened — but not whether that action was acceptable.

With traditional software, an execution usually follows a path defined by the developer. Agents are different. They can reason, choose tools, change their path and take actions we didn’t explicitly define.

That creates a new infrastructure problem for teams deploying agents in production:

What did the agent do? Why did it do it? Was it allowed to? Which policy and permissions applied? And can we prove what happened afterwards?

Traccia is our AI Agent Control Plane — a vendor-neutral layer to observe what agents do, evaluate how they behave, govern what they’re allowed to do, and audit what happened.

We’ve open-sourced the Traccia SDK and built it developer-first, with OpenTelemetry at the foundation. It works across models and agent frameworks, so teams can add observability and governance without being locked into a single AI vendor. We’re still early, and we’re building this alongside developers and teams actually deploying agents.

I’d especially love feedback from people running agents in production:

What are you using today to debug agent behaviour?
How are you evaluating agents?
And more importantly — how do you control what an agent is allowed to do?

We’re also making it easier to try Traccia during the launch.

3 months free with coupon code: TRACCIAPH

Hey Product Hunt 👋 We’re live.

I’m one of the makers of Traccia. If you’ve ever watched an agent take a tool call you didn’t expect and then scrolled a 2,000-span trace trying to answer “was that even allowed?” - that’s the pain that started this.

What ships today

• Open-source SDK (Python + Node), OpenTelemetry-native

• Full-fidelity traces across models/agent frameworks

• Eval path: prompts → datasets → scorers → experiments before you promote

• Runtime governance: policies + evidence so “observe” isn’t the end of the story

Who this is for
Teams putting agents in production - not demos. If your stack already tells you what happened, but not whether it should have happened, you’re our ICP.

One ask
If you run agents in prod, comment with your current stack for:

1. debugging a bad tool call

2. deciding promote vs rollback

3. blocking an action at runtime

Even “we use X and it’s fine/it sucks because Y” helps more than a silent upvote.

Trying Traccia today? Platform is open - use coupon TRACCIAPH for 3 months free. I’ll be in the comments all day and will answer everything personally.

https://traccia.ai

- Aditya (and the Traccia team)

The Tsunami of AI agents will overhelm the industry. Every organization and team wants to build AI agents. Yet, once the initial excitement fades and these agents are operating at scale, difficult questions will emerge

• Was my agent actually allowed to do that?

• Why did it make that decision?

• Could I have prevented that action?

• How do I stop a rogue execution before it causes harm?

Traccia as an AI agent control plane ensures you have the right visibility and control over what you agents can do, what they can access, terminate calls on policy violation and many more.

Trying Traccia today? Platform is open - use coupon TRACCIAPH for 3 months free

Congrats on the launch. Open sourcing the SDK is a lot to give away for a control plane, good to see.

Did there occur any specific cases with AI agents that you could label as anomaly within their process?

Having one place to manage AI agents across different platforms is huge, especially now that everyone's using a mix of frameworks. Really solid build!

Congratulations on the launch! It sounds like a really useful project based on the description. I’ll show it to our team - we’re currently using a similar solution...

Great control & observability layer guys, solid work. Love the fact there's no vendor lock-in.🔥

Vendor neutrality is important. How easy to setup and commission?

Hey Product Hunt 👋

I'm Mingtian, co-founder of PageIndex.

PageIndex lets you ask across your entire document set and verify every answer down to the source line. If you work with long and professional documents where accuracy and traceability matter, financial reports, legal contracts, textbooks, research papers, etc. and you can't afford to trust an answer you haven't checked, this is for you.

✍️ Here's how it works:

1. Drop in your entire folder. Folder structure is preserved, and everything you add stays in your knowledge base, so you build it once instead of re-uploading the same files into every new chat.

2. Ask your hardest question across all of it. Ask the one you actually need: a number buried in an appendix, a clause that only makes sense with the definition twelve pages back, a figure that has to be pulled from a table and compared across four files. PageIndex goes and gets all of it.

3. Verify answers in one click. Every answer comes with micro-citations. Click one and the source document opens beside your chat at the right page, highlighted at the exact line the number came from.

Why you want PageIndex:

• Checking a number takes one click instead of an afternoon of opening PDFs

• One question runs across hundreds of documents, not one file per chat

• Tables and charts are read in context

• Your library compounds each quarter instead of starting from an empty chat

Leading accuracy on FinanceBench. 30K+ people use it, and the retrieval engine underneath has 35K+ GitHub stars and hit #1 on GitHub Trending.

🎉 Launch offer: code PRODUCTHUNT gets you one month of Pro free.

👉 app.pageindex.ai

Thanks for checking us out. I'll be here all day.

Hiii PH, I'm Cathy, GTM at PageIndex.

I'm the non-engineer on this team, which makes me test subject number one. If I can't get a verified answer out of a folder in thirty seconds, it goes back to Ray.

Before this I worked in finance. I studied it at LSE and ground through CFA Level I. None of that helps you at 11pm when you're pulling the inputs for an EBITDA build, then tracing every single number back to the page it came from.

👇 That's the bit that changed for me:

• My filings sit in one folder that stays indexed.

• I ask across all of them at once.

• Every number in the answer carries a reference. One click lands me on the exact highlighted line, right next to the chat.

• The verification pass that used to take an afternoon is now one click.

One favour if you're testing it: please don't ask what Apple earned last year. Any chatbot answers that. Ask the thing only you would know where to look for:

• a covenant threshold buried in an appendix

• a segment number that moved between restatements

• a figure that only appears in a footnote

If anything confuses you in the first two minutes, tell me bluntly. That's the feedback I act on fastest. I'm in the comments all day ☕

For contacts and reports I'd rather spend a few seconds checking a source than blindly trust a summary. Does it preserve the orignal document formatting when I kump to the citation?

Asked three follow-up questions in a row and it kept the context each time without me re-explaining anything.

Hi PH, Ray here, co-founder and CTO of PageIndex.

I did my PhD in databases at Oxford. Years spent on indexing, and I was firmly on the side of vector databases.

I've changed my mind about them being the right infrastructure for retrieval in AI systems. I want to be precise, because the claim is not "vectors are dead".

An index is defined by the one question it can answer. A vector index answers: which chunks look most like this query?

• Right question for broad recall over messy, conversational text. Vectors will keep doing that job well.

• Wrong question for a long, professional document.

Two reasons it breaks down there:

• The passage you need may share almost no wording with how you asked for it.

• The answer often sits behind a cross-reference like "see Appendix G". No amount of similarity gets you through that pointer.

So we index the structure instead of the surface.

• The document's tree sits inside the model's reasoning context.

• It decides where to look next, not what looks similar.

• Retrieval becomes navigation. Navigation leaves a path, which is why every answer can point back at the source line.

If you want to stress-test it: upload the longest and most complex document you own, then tell me what happens. Bug reports are worth more to me today than upvotes.

Congrats! Avoiding embeddings is an ambitious architectural choice. I'm especially curious whether this makes the results more interpretable, since each search decision could potentially be traced through the document tree.

Feels like what Perplexity does for the web, but pointed at my own document library instead.Great Launch!

I like the exact-line citations. Congrats!

Can I share a single answer with its citations as a link, so a colleague can see the sources without an account?

Congrats on launching PageIndex! Verifying answers with clickable citations solves a real pain point. How do you handle documents with scanned or image based pages?

This seems especially valuable for investors and researchers.

The document should always be the source of truth. I like that AI is supporting the reading process instead of replacing it.

can you suggest how pageindex is better than chatgpt or claude code. what makes pageindex better?

Looks pretty scientific :) Do you collaborate with some universities and research centres? :)

Hi Product Hunt. I'm Elina, and this is my second launch. 🐱

Yomi is a little cat who loves being read to. Kids read stories out loud, Yomi listens (speech recognition), and every word feeds the cat a little. Tomorrow Yomi is hungry again.

I'm a designer, and I've spent years building educational tools with teachers. In a workshop with first- and second-grade teachers, one question kept coming back: why do some kids freeze when asked to read aloud, even when they can read just fine on their own? A few months later my own kid was doing exactly that, reading silently, refusing to read aloud to anyone. I didn't want to force it. I wanted to build her somewhere safe to practise.

The apps I tried first made it worse, not better: streaks that made bedtime stressful, fire emojis that made her feel bad for taking a day off.

So Yomi deliberately has:

– no streaks (nobody should lose a 14-day streak at bedtime)
– no ads
– no accounts, no data collection
– no leaderboards, no coins, no levels for the kid, only the cat, who is happy today because you read to it

To be honest about what it is: Yomi won't teach your kid to read. Teachers and parents do that. Yomi just gives kids a reason to want the practice, real stories at three levels, tap a word to hear it, and a little wizard for making up your own stories to read back.

It's on iOS in English and Finnish. I'd love to hear from parents here: what made reading practice work (or not work) at your house?

oh i love this idea and style so much! please add German as supported Language <3

This is genuinely adorable haha. I can totally see kids getting weirdly invested in making sure the little cat gets fed every night. Such a fun way to make reading feel less like homework. Congrats on the launch! @elina_patjas

My kid isn't ready for this yet but I love the heart behind it!

I've been thinking (and reading) a lot about the literacy crisis recently, so this is well-timed. I can't wait to check it out and introduce it to my nephews! Congrats on your launch 🎉

Loved the idea Elina! Wish you all the best here

This is such a cute app! My nephew is going to love this!

@elina_patjas This is going to end up being the first kids focused app I've seen launch here that I actually download and use with my son. I really love the lack of gamification in this. Like you said, my kid does not need the stress of trying to keep a streak alive when it's bedtime haha.

To your question of what's worked: my son really likes when I ask follow up questions after we read. And from my side, I've found that it really helps to actually get him engaged with what we're reading when he knows we're going to talk about it afterwards. So there being comprehension checks in this is going to be a big hit.

I really appreciate seeing something like this that's actually useful for kids and doesn't have a bunch of bs packed into it. Excited to try it out!

Great idea for an app! Any technology that helps kids with reading and getting practice pronouncing words is super helpful. Some ideas for future iterations could be adding pictures to the books somehow, or possibly having Yomi meow/say some phrases to come alive even more. Keep up the great work!

Love this cool app! I have 4 kids and I am sure kids will love this app

Hi Elina, I'm not a parent, yet, but I love this. The design and graphics are amazing. Good luck with the launch.

the "file tree mirrors the disk, not the system" line is a good way to put it - most context tools are really just fancier grep. what I use today is basically Claude Code's own repo-map plus whatever I paste in manually, and it breaks exactly where you'd expect: cross-repo calls, since there's no single view spanning our services. the part I'd want to understand before adopting this though is graph freshness - on a fast-moving repo with agents committing constantly, is the graph updated incrementally per-commit, or is there a reindex lag where an agent could query stale call/import data right after a merge?

Congrats on the launch folks!!

GitNexus feels like a practical solution to one of the biggest bottlenecks in agentic coding i.e. the context. Huge congrats on the launch, Subham!

45k stars is wild 🤯 deterministic graph over embeddings just makes sense

This is super cool!

At Flowtogen we are using GitNexus and it is helping our team.

45K GitHub stars - that’s impressive! Is there a mechanism for controlling who can access the code? Often you can’t show the entire codebase to everyone, especially for financial applications.

Hi Product Hunt,

SpacebarX is an offline notes-and-tasks workspace that keeps your work on your machine.

I built it because I was constantly splitting my brain between two apps: notes in one, tasks in the other, and the context I needed always somewhere else.

With SpacebarX, everything lives in one nested outline. Start with a thought. If it becomes a task, make it a task. No project setup, no predefined templates, no deciding how to organize before you can begin.

It has the simplicity of an outline, with the focus and efficiency of a task manager, while keeping the notes behind each task close at hand.

The part I care about most is ownership:

• Your files stay local

• You can work fully offline

• If you want sync, connect your own Google Drive or Dropbox

• There is no SpacebarX server storing, reading, or controlling your notes

A few things you can use it for:

• Notes: Keep decisions next to the project plan they came from.

• Tasks: Turn any thought into a task, assign due date to it and see the task in Today / Calendar view.

• Files and folders: Separate work into files and folders when you need structure, then switch to Whole View to see it all as one connected outline.

• Search: Find anything without losing the surrounding outline and context.

• Writing and code: Keep Markdown and code blocks beside the work they support.

• Planning: Use Board, Table, Mind Map, or Calendar views when an outline is not the best way to see the work.

Right now, SpacebarX is a PWA you can install from your browser on desktop or mobile. Native desktop and mobile apps are in progress.

The free tier is free forever. Pro is available monthly, yearly, or as a $99 Lifetime plan, normally $199. It adds Board, Table, Mind Map, Calendar, saved searches, encrypted documents, version restore, uploads, and bring-your-own-key AI.

If you have tried combining notes and tasks in one app and gave up, I'd genuinely love to know where it broke down for you?

Hey Product Hunt 👋

I'm Jason, one of the makers of Caddi. Caddi is an agent that builds agents.

We build for law firms, RIAs and accounting firms, and the same thing kept coming up: the work isn't hard, it's just endless. A contract comes back signed, someone downloads it, renames it to the firm convention, files it to the matter, logs it in the CRM. Forty times a month. Every month.

Automating that has always meant a project. Scope it, spec it, configure it, get a developer. So it never happens, and people keep doing it by hand.

Caddi takes a different path, and it starts before you build anything: it reads your stack and tells you which work you repeat most, ranked by impact.

You pick one, and then you teach it like a new hire, except this new hire already knows the job. It follows your guidance or guides you with best practices, and probes for edge-cases: not just "download the contract", but "say how you tell an executed copy from a draft".

When it hits something ambiguous it stops and asks, with the options laid out and a confirm button. What if only one party has signed? Your answer doesn't disappear into a transcript. It becomes a rule in the agent.

What comes out reasons with AI where judgment is needed and runs deterministic code where it isn't, so the messy cases get handled without the exact parts going off-script. Then it runs in the cloud, executing thousands of actions at a time.

And you can see all of it. Every run is a log of what it decided, step by step, with the permission it used to do it: read, create, send.

We'd genuinely love your feedback, especially on where you'd trust an agent and where you wouldn't. That line is the whole product for us.

I can see this being useful for all those boring admin tasks we keep doing every day.

ui looks very interesting.

How does it handle small change in the process? That's usually where automations break for me.

I like the screenshare approach. Feel more natural than setting up a workflow step by step.

what happens if a website changes its layout slightly? Does it adjust or break?

Talking out loud while doing the task feels way more natural than manually building a giant diagram.

Can it handle workflows that hop across 3 or 4 different apps in one go?

How well does Caddi handle exceptions when the real process differs slightly from the example?

Hey Product Hunt 👋

I'm Owen, founder of Kira.

For most of its life, Kira has been a tool. You bring a photo or an idea, and Kira turns it into something beautiful: a portrait, a cinematic shot, a video. I've always loved that.

But one thing kept nagging at me: the second someone made something they loved, they wanted to show it to someone. The making was never really the end of it.

So today we're launching Kira Community: the biggest thing we've ever shipped, and our first real step from a tool into a place.

Here's the idea. Every creation becomes a moment you can curate, caption, and tag: #portrait, #goldenhour, #summer, whatever it is. Your moments live on your profile, and they flow into feeds you explore by hashtag, so instead of scrolling past strangers, you find the people who are into exactly the same things you are.

We spent years making it easy to create. This is the start of the harder, more interesting part: making it worth sharing, giving every creation an audience, and every creator their people.

"Make joy worth sharing" is where we started. Today it becomes something bigger: make #anything worth sharing.

I'll be in the comments all day. Would love to hear your thoughts (and what you'd want to see next).

— Owen

Congrats on the launch! i tried to access the site, but it says i'm blocked. is philippines blocked? im from the philippines

Looks great! What does the free tier cover? Would love to try it on a couple of real projects before committing.

I just did this and was super super surprised at all the wrong things that AI said about me! Definitely going to show this to my coworkers at Optiver!

Super pumped to see everyone interacting with Pluto! We're excited for everyone to finally have their professional career represented by Pluto!

Super excited for everyone to see this! Please let us know any feedback and comments. We are always building and love any kind of feedback!

Big congrats Sahil! 🚀 Pluto is such an interesting take on how AI will change professional discovery. Excited to see this one grow — all the best for the launch! 👏

Been really exciting watching Pluto come together and even more exciting to finally share it. We would love any feedback, ideas, or thoughts as you try it out 🚀❤️

Let's gooo

very cool launch! Congrats team

Super interesting @sahilseth!
One thing i learned testing out products like mercor, clera & jack&jill it's super frustrating to jump in the intro calls with each agent seperately. A partnership to share the context from Pluto with these automatically, would be amazing and save a lot of time.

Super excited about this launch. I’ve been following Pluto for a while now and am really interested in seeing how it changes the way companies find and reach great talent! Congrats to the team, it’s only up from here 🫡

Congrats! Being discoverable by agents is huge. LinkedIn has become pretty spammy recently

Jenny's comment right above is the whole risk in one line. A resume is thin, but I wrote it, so when it's wrong that's on me. A profile inferred from a 10 minute call is wrong in ways I never said and probably won't think to check, and it's the version an agent reads when I'm not in the room. Control only means something here if I can see which sentence came from which thing I said and kill it at that level, not a settings page.

Intriguing idea! How does it work from the hiring person's perspective to interact with the candidate's agent?

Super interesting and a great way to expand my network!

Hi Product Hunt! I'm Angelica, one of the makers of Canvas!

Canvas is part of Wondering Desktop Web and is a visual way to understand things in parallel. We built Wondering because we felt like AI was giving us more answers, but we are understanding less.

When we want to learn a new concept for work, unpack a research paper, or understand any new topic, the learning process is rarely linear. Yet, when we ask AI, we can only ask follow-ups in the same chat thread or create new conversations so we don’t clutter up the first.

Canvas changes that!

Features:
🖼️ Replies with visuals and interactive diagrams, not just a wall of text
Every response tries to show, not only tell. We add visual explanations with interactive diagrams so the information becomes easier to understand.

🌱 Spin up new chat threads
Whenever you find a term you don’t understand or sentence to follow-up on, you can spin it off into its own thread without cluttering the original conversation.

💬 Suggestions galore
Sometimes you don’t know what you don’t know. We highlight all the key terms so you can dive deeper into understanding those building blocks. We also suggest relevant questions you can ask in the chat thread or spin up a new one.

🧹Organize related chat threads together
All related explorations are now grouped into one Canvas so it’s easy to revisit and make connections.

🗂️ Context stays together
No more copy-pasting answers from one chat thread to another. Creating a new node from the first already pulls its context.

📝 Classic highlights and note-taking
Of course, we have the ability to classically highlight and add notes to the whole chat thread or to a highlighted section of the chat.

Our goal is to make AI feel less like an answer machine and more like a learning environment, where you can build understanding one step at a time.

You can try it out at https://wondering.app/canvas

If you try it, tell me what you’re using it to understand and your thoughts. I’ll be here all day.

This was awesome to try, discovered on X btw

Can teams together collaborate on this?

Can I use it for wireframing website as well?

Branching is the easy half. Every canvas tool I've tried is good at spinning threads off and useless at bringing them back, so you end up with a dozen branches and no way to say which three actually changed what you understood. If a branch can't fold back into the parent as a conclusion rather than a transcript, the canvas turns into a graveyard by the third session. Worth watching the cost side too, since each thread carries its own context and exploring gets expensive faster than it feels like it should.

There is an abundance of choice when it comes to text editors, but they all tend to fall into one of two categories. Either their interface is too cluttered and bursting with features – or they are too simple and limiting.

With Kraa, we tried to strike the best balance between the two. Create an editor that's pleasant to use, simple, clutter-free. And yet – it can take anything you throw at it.

This update also includes a voice control feature unlike anything you have used before.

Kraa is built on ProseMirror (via TipTap) and Svelte.

You don’t need an account to try it. We are proud of the improvements in this new version and hope that you will enjoy them!

Live demos (no login required):

• Chat

• Blog article

• Long-form story

• Magazine

Every blog post I publish started as a note in a different app, and the handoff is where the draft usually dies. Which of those jobs do people actually stop using another tool for?

Hey Product Hunt 👋

Every AI meeting tool I tried joins the call and mails you a summary afterwards. That helps the archive, not the conversation. What I needed was help during the call, and the answer was usually sitting in a document I had already read twice.

So I built Savvy.

Point it at a folder of documents and it builds a versioned brief per client. During the meeting it stays quiet. It speaks up for exactly three reasons:

1. They asked something your brief already answers.

2. Someone crossed a red line you set.

3. You pressed Advice.

Every card cites the document it came from, so you can check it before you say it.

It runs on your Mac. Files, indexes and transcripts stay there; only excerpts and the audio stream leave it. Apple Silicon, macOS 13+. Open source, MIT: https://github.com/jamalavedra/savvy

If you take client calls, what do you keep open on the second screen today?

Hi everyone!

Qwen has done this once before. Qwen3-Next gave everyone an early look at the architecture that later showed up in Qwen3.5.

Qwen3.8-Flash-Next is doing the same for Qwen4.

It’s a 125B model with only 6B active parameters, and a lot of the new design is about getting more capability without dragging compute up with it. QSA makes long-context retrieval cheaper, Gated Residual gives information more paths through the network, and the new N-gram memory adds capacity with very little per-token compute.

Qwen keeps putting these architecture previews out as real models people can actually run, well before the next generation arrives.

And the weights are already up :) (with a license worth reading)

the gated residual + n-gram memory combo is the part i'd want to poke at. changing how information flows through the network usually means old LoRA adapters and fine-tunes built for Qwen3 don't transfer cleanly to the new architecture. is that the tradeoff here, or did you design it so existing Qwen3 tooling and adapters still mostly work on top of this

Congrats on the launch! love that these previews are actual downloadable models.

the active param count is spot on — 125B total is what kills it for local though. any chance of a ~35-40B/A3-6B version on this architecture? 35B-A3B was the local goat and it has no successor yet 🙏

Hi Product Hunt. I am Mio, founder of Ojin.

I have sat through a lot of AI demos where the face looks perfect but the conversation is unstable.

You say something, it waits. You pause to think, it talks over you. You interrupt, and it finishes its sentence anyway.

Everyone nods and nobody says the obvious thing, which is that this is not a conversation.

So we built Human AI Agents.

What it is

One still photo, a persona written in plain language, and a voice. You get an agent you can talk to and interrupt.

What runs underneath

Two face models behind a single API. Portrait for speed and scale, Presence for expressiveness. Both stream over WebSocket and drop into Pipecat or LiveKit.

What we actually spent the time on

Turn-taking. Knowing when someone has finished a sentence rather than paused to think. It is the unglamorous part and it is most of the product.

Try to break it. Interrupt it mid-sentence, talk over it, trail off, use an accent, most of all have fun!

I'll be around all day.

Congrats on the launch! ✨ Got some nice recommendations from Sofia for travelling and it continues the discussions pretty smoothly when interrupted or when I interject something randomly (and knows to respond to that also). It still is obvious that it's an AI, but the experience is pretty good.

So happy to see Ojin out today after all the hard work! Proud to work with this team 🔥

Proud and honored to be a part of the Ojin team. Hard work do pay off. Congratulations to everyone at Ojin. My favourite is Presence. 🤩

I'm very excited to finally have Ojin released!! A lot of engineering efforts went into making this product possible and I'm so proud of the team. Can't wait for people to try it out!

Happy to have contributed to this as a product engineer. It's been a wild few months of building, and it's great to finally see people trying what we made. Huge respect to the team, genuinely talented and great people to build with. Let's go! 🚀

Congrats on the launch! Had a lot of fun with the API portal so far.

We​'re ​v​ery excited to ​s​hare Ojin today. ​P​lease check it out and share your feedback, thoughts​ and questions​. We're eager to hear what you think - the more feedback, the better!

Real-time voice makes $/minute tempting, but I’d rather know cost per completed call outcome. A 4-minute call that books the meeting can be cheaper than a 60-second call that goes nowhere.

Turn-taking being most of the product is the right admission, and it's also the thing nobody buys on. We work on the generated side of avatars and the failure mode is different, a face that holds for eight seconds and falls apart on the ninth, but the rule is the same, the boring temporal part is where the work actually is. I'd put endpointing latency on the page next to the face model specs, because two face models behind one API reads like the differentiator and it isn't. Anyone who has shipped voice will pick you on the interrupt handling.

Congrats on the launch! I use LemonSlice right now but am always open to new providers. Do you have a differentiator?

Hey Sendra team!

I’m actually converting my emails from Figma to HTML right now, and this tool helped me figure out how to do it without dealing with code, plugins, or all the other technical stuff — definitely not my area of expertise.

Thanks, guys 🙌

just left a review, the test-to-inbox flow is genuinely the best part. saw your reply about gmail dark mode and outlook color shifts still being the rough edges. curious if sendra flags those as "known risk, check manually" before you even send the test, or if you only find out once the test email actually lands and you're eyeballing it yourself

I design emails in Figma and then something always breaks in one client I didn't test. Which one still breaks last for you?

What a nice find out of the sea of AI agent launches. This made me smile. Congrats on the launch!

to answer your actual question - yes, the certificate made me laugh, "Department of Avian Emotional Welfare" is exactly the right amount of fake-official. one thing I'm curious about since it's built as a gift: if someone fat-fingers the name at checkout (typo, autocorrect, whatever), is there any way to fix it after paying, or does the duck stay permanently happy under a permanently misspelled name? feels like it could go either way as a joke - "even the mistake is now official" is funny too, but I'd want to know before I gift one to someone.

Hahahaha i'm enlightened, there is a new path from 0 to 1M$