PH热榜 | 2026-03-18

AI tools stop working when you leave your desk. Dispatch doesn't.

It's a feature inside Claude Cowork. Assign tasks from your phone. Claude executes on your desktop. Come back to finished work.

What it does:

• 📱 Message Claude from your phone

• 🖥 Claude accesses your files, browser, local tools

• 🔒 Sandboxed. Local. You approve before it acts

• 📋 Returns reports, tasks, actual output

Use cases:

• Reports from internal dashboards

• Finding better flights

• Anything Claude can do on your desktop, from anywhere

Good to know:

Desktop must be on. Max subscribers now. Pro in days. Research preview, more coming soon.

Try it. What's the first task you'd dispatch?

I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

This is a much-needed development. I appreciate OpenClaw's impact on the industry, particularly its most impressive feature: remote control. Now, features like Claude Code's remote control and Claude Dispatch are bringing this same remote capability to autonomous systems.

I believe the logical next step will be integrating these tools into messaging applications like Slack, Microsoft Teams, and WhatsApp. We can already see this potential today, as collaborating with Devin on coding tasks directly within Slack is extremely convenient and time-saving.

Love the ‘desktop as an engine’ approach. There’s always that 10-minute gap between leaving the office and getting home where a random task pops into my head :)

Just love anything anthropic/claude related. Claude itself has done so much more then chatgpt ever could.

Is this similar to the /remote-control feature inside claude code?

this is a cool idea honestly. being able to kick off something from your phone and have it run on your desktop makes a lot of sense, specially for those random moments when you’re away but still wanna get something done.

curious, what are people using it for first mostly, quick research tasks or proper work stuff like reports?

Congrats on launching! Texting Claude a task while you're out and coming back to finished work is exactly how AI should fit into out days. Does it maintain context across multiple dispatches or does each message start fresh?

This has been something I've been waiting for! For those quick tasks you want to schedule on the go..

love it. congrats guys.

This is like remote controller for AI

Amazing ! A Claude Cowork for Linux coming soon ?

Hey Product Hunt — I'm Keith, co-founder of Lightfield. Before this, I led Instagram Direct from zero to 500M users and helped build Stories and camera AR. I then co-founded Tome, which launched as #1 on Product Hunt and grew to 20M users.

Of all the products I've worked on, Lightfield has the deepest engagement I've ever seen. It’s an AI-native CRM that builds and updates itself from your real conversations — emails, calls, and meetings. No fields to define. No manual data entry, ever.

Connect your inbox, upload data from a spreadsheet or your old CRM, and you're up and running in five minutes.

One of our early users connected her inbox, typed one prompt — "go through my emails and fill in my opportunities" — and came back to a fully populated pipeline. Stages, contacts, deal context. All built from conversations she'd already had.

Most AI CRMs stop there. Lightfield goes further — it stores every interaction as continuous context, not disconnected records, so the AI sees the full story of every relationship. Then it writes and runs code against that complete memory.

Ask it anything:

• "What objections keep coming up across all my calls?" — pattern-matched across every conversation you've ever had

• "How has our ICP shifted in the last 3 months?" — answered from real deal data, not gut feel

Then put it to work:

• "Draft a re-engagement email for each of my cold prospects" — written with full knowledge of your history together

• "Build a board-ready pipeline report" — from what's actually happening inside deal

We built this for founders doing their own sales, from our own experience. That's where every company's customer knowledge begins — with the founding team and their first conversations. Lightfield makes sure none of it gets lost. Not when you hire your first rep. Not when you build out a team. Every conversation compounds.

As a thank-you to the Product Hunt community — we're offering 3 months free for anyone who signs up with code PH3. We ask for credit card up front at sign-up to keep bot traffic out, so don’t forget to use the code to get free access.

Try it at https://lightfield.app

Our entire team is on standby to help you get set up. Drop a comment with your first question for the agent — I'll reply to every one.

Huge thanks to @chrismessina for hunting us. We're honored to have the #1 Product Hunter behind our launch.

Every CRM I've tried has the same failure mode: it's only as good as what people remember to put into it, and nobody does consistently.

The auto-capture from emails and calendar is the piece that actually solves the behaviour problem rather than just the interface problem. Most CRM redesigns just make it easier to enter data manually. This skips the entry step entirely, which is a different category of solution.

Running a SaaS and managing a mix of inbound leads and partnership conversations, so this is directly relevant. The question I'd have is around signal quality: when the AI is reconstructing deal context from email threads, how does it handle relationships that are mostly async and fragmented across months? That's usually where CRM data gets thin and unreliable, regardless of the tool. Congrats on the launch!

nice one. a crm that actually stays updated on its own sounds way more realistic than asking sales teams to log everything properly forever. the part about asking questions from actual conversations is pretty interesting too.

curious, what kind of teams are adopting it fastest right now, smaller sales teams or bigger ones with messy existing crm data?

this is a pretty strong pitch honestly. the zero manual data entry part is what stands out most, because keeping crm data updated is where a lot of teams quietly lose consistency. also like that it is not just storing info but actually helping teams do something with it after.

curious, when people switch from hubspot, what’s the first thing that usually makes them feel the difference most?

When I heard Lightfield's founders shut down a product with 20M users to build this, I knew it had to be pretty special. And, after a couple months of early access I totally understand why they did it. I've never been good at updating my CRM. But having it updated is actually super valuable... it's just the kind of tedious work I either feel too busy to do or just forget. Lightfield is the first CRM I've used where I feel like I'm getting that value without having to do the work myself. It's definitely the first CRM I've actually loved using.

Love the design and approach; proven team.

I appreciate Lightfield's proactivity as well — once I connected my contacts and calendar, it started offering value on Day 1 with meeting pre-caps and client dossiers.

I was one of the very first people to sell Contact Managers and then Sales Force Automation products to salespeople in the late 80s. The salespeople hated it because of all the input it required... among other things.

CRM has come a long way since then. You've got a good start. Once you add Customer Service and Marketing you'll truly have an integrated AI-based CRM solution.

The biggest CRM problem I have seen is not the setup — it is that people stop updating it after a few weeks. Zero manual entry sounds like it solves that. But what happens when part of the conversation happens on channels you are not connected to — do you end up with gaps that make the data unreliable?

that looks amazing :) I always struggled to force sales people to keep up and fill out CRM data properly. More often than not, the end solution was to set up a lot of guardrails, firing out warning messages every time somebody forgot to follow the expected flows.
I'm wondering though, how customizable is this? because different customers will likely have different pipelines!

This is exactly the right wedge.
Most CRM tools still depend on humans being disciplined data entry machines which almost never scales in real life.
Auto-capturing context from email + calendar is a much stronger behavior-layer solution, not just a UI improvement. Smart product. Congrats on the launch 🚀

Excited for this launch! Congrats team

Lightfield is one of the best products we’ve used—especially valuable for transitioning from founder-led sales to scalable growth. Congratulations to the team for the PH launch. If anyone is thinking to migrate from Hubspot or other CRM, please go for it. Lightfield will make sure that you will not miss those CRMs.

Got a chance to speak with their founder, Keith, at a founder mastermind and was pretty blown away with the vision and capabilities of the platform. Huge congrats on the launch.

Congrats on the launch team!

@keith_peiris mentioned that Lightfield stores interactions as continuous context and the AI writes and runs code against it. That sounds super interesting! Curious how that actually works in practice. When the AI drafts a re-engagement email, is it pulling from a different slice of that context than when it answers 'what objections keep coming up across deals'?

And as the team grows from founder-doing-sales to first few reps, does that accumulated context become a shared asset everyone can query, or is there a way to scope what's visible?

This is a great take on CRM.

The biggest problem was never the tools it was getting people to actually update them.

If the system builds itself from real activity, that’s a huge shift.

Curious how you handle accuracy and edge cases.

Love the pitch CRM that builds itself” is instantly compelling. Could it even harder by leading with the key outcome: “Stop manual data entry and get actionable insights for your real conversations in minutes.

Congrats ! It’s a great great product. Features in my top 3 products!

I could not believe first time when I saw the all data

auto populated back to 12 months !the context it pulls from Emails / calls and intelligence on accounts and people is exceptional! Our entire team is on it!

Great product. Congrats on the launch!

Congrats on the launch Keith! Rooting for y'all!

I love Lightfield and what they are building. Strong upvote from me!

Obsessed with this!! Can't wait to try it out, was literally wondering how hard it would be to train my own agent to do this. Glad I don't have to.

This seems super handy! Is it meant to replace tools like HubSpot or Klaviyo, or does it work alongside them? Also, can it connect with support tools to help build a CRM list?

Way to go! Lightfield is so well crafted.

@Lightfield is legit. I talk to my CRM every day now. Would've sounded insane a year ago.

We've been through probably 5 CRMs over the past few years...none of them stuck. Too much manual input, too many unnecessary features.

Lightfield captures details from every call, email, deal, and actually makes the context insightful without much input. We're a lean team that moves fast. Lightfield is a big reason prospects don't slip through the cracks anymore.

WOW, sounds to be the next generation of CRM. Are there a way I can import the data from my previous CRM ?

Congrats on #2 Lightfield team, way to go!

Hi Product Hunt! 👋

I'm Davorin from the Databox team, and today we’re excited to share something we’ve been working on for a while: Genie, your AI analyst.

Most teams today have dashboards and reports that track everything from revenue to marketing performance. But when someone asks a simple question about what’s actually happening, like why revenue dropped or what’s driving the growth, getting clear answers is still harder than it should be.

You have to dig through multiple dashboards, compare metrics, identify trends, or even loop your data team. And by the time you finally have a clear explanation, you’ve already burned hours you didn’t really have.

We kept seeing the same pattern: even though teams had plenty of data, they were still struggling to get answers fast enough to make decisions.

So we built Genie. 

Genie is an AI analyst built directly into Databox. Ask questions about your business performance in plain language and get answers instantly, with trends, visualizations, and context drawn directly from your metrics.

Here are a few things you can do with Genie today:

🔎 Ask questions about your data: Understand why metrics changed, uncover insights faster, or dig deeper into what’s actually driving performance. 

📊 Create metrics and dashboards: Describe what you want to build, and Genie can generate the metric or dashboard for you.

💬 Share analyses instantly: Send Genie conversations to anyone so they can see the analysis and context behind the answer.

🪄 Chat with Genie from your AI tools using MCP: Explore performance and uncover insights directly in tools like Claude or ChatGPT. 

Genie is designed to help anyone on your team explore and understand performance data, so decisions don’t have to wait. 

-----

We’d love your honest feedback: what’s the #1 question about your business data you wish you could answer instantly?

Drop it in the comments - we’re building Genie with teams like yours in mind, and your feedback will directly shape what we work on next.

Thanks for checking out Genie. Every comment, idea, and upvote means a lot to our small team. 🙏

This is awesome and a very clear peek into where business analytics is moving in the short term. Congrats team!

One question: is there already a Slack integration or potentially coming soon? @davorin

As a Dutch online marketing agency (Unison), we used Databox for client reporting for years and recently tested alternatives like GoMarble, AdSuperpowers and Claude MCPs. Having had the opportunity to beta test Genie, we can confidently say it easily beats them all. In a short time it has transformed how we work.

The standout feature for us is how data is strictly scoped to a specific 'client space.' When we query data, there is absolutely no risk of it mixing with other clients' data sources. This eliminates noise and hallucinations completely, which is a massive advantage over competitors.

Because the platform is so intuitive and accessible, the adoption within our team has been incredibly fast and widespread. Genie instantly became a core part of our daily workflows. In fact, whenever 'account maintenance' is on our schedule, using Genie is now a standard step in that process. It makes combining data sources and analyzing large datasets effortless. Drawing in-depth conclusions is so much easier now, mainly because you can literally just ask the platform the exact question you have in mind.

At Unison, we strongly believe in a balanced collaboration with AI, where a Human-in-the-Loop (HITL) remains a crucial part of the process. Genie perfectly facilitates this philosophy. It does the heavy lifting with the data, but keeps our specialists in control to validate and act on the insights. Highly recommended for any agency looking to level up their data analysis.

Congrats on the launch, Databox team!

this is pretty cool tbh. a lot of teams want insights from their data, but the setup and learning curve usually slow everything down. having an ai analyst inside databox feels like a smart way to make analytics more approachable.

curious, are people using genie more for quick checks or deeper business analysis?

This looks interesting, especially for teams that want answers quickly without having to build everything from scratch first. The idea of having an AI analyst built right into the BI workflow makes a lot of sense if it can actually help people get to insights faster.

Curious, what kinds of questions are users asking Genie the most right now?

What excites me about Genie is how quickly it changes your relationship with data.

As a Director of Engineering at Databox, a big part of my job is turning metrics into context. Reports, updates, explaining what’s actually going on behind the numbers. Genie shifts that from “looking at dashboards” to actually talking to the data.

And from an engineering perspective, that “just works” feeling is anything but simple. You’re dealing with cross-data-source querying, interpreting intent, and matching that to the right data in real time. When it feels effortless, it usually means a lot of hard problems were solved behind the scenes.

The practical impact is immediate. Less time navigating charts, more time understanding what’s happening and communicating it clearly. I get that time back to focus on actual impact.

It’s one of those shifts that feels obvious once you use it.

I saw a post on linkedin that you were coming out with a new chatbot so decided to give it a shot.
To be honest it is one of the cleanest and well chatbots I've seen in a long time. It especially surprised me with the diagrams it made and just how well it worked for just coming out! Good job!

Super proud to be part of the team behind this.
As a sales consultant, I’m always careful about what I recommend vs. what I actually use. Genie is both.
Use it every week, not just in demos 😄
Huge congrats to everyone who made this happen.

I couldn’t find this on the landing page, so asking here... how do you ensure that data we upload or connect remains fully private? What security standards are you following?

Congrats on the launch.

Huge congrats to the Databox team on launching this! 📈

I’ve been playing around with the beta for a week leading up to today, and the value it provides is immediate. It’s seriously useful for cutting through the data noise. I really love how it drafts short, contextualized reports and quick visualizations. I even prompted it to turn those insights into complete dashboards that I can send directly to stakeholders on Slack or via email! It perfectly bridges the gap between staring at metrics and actually understanding the insights. Fantastic addition to the platform!

As a data analyst, I was genuinely curious how Genie would handle nuanced questions. I tested things like month-over-month retention by segment and root-cause questions about churn spikes. The answers were accurate, and the reasoning was sound. It will not replace deep analysis - but for the 80% of everyday data questions, it delivers consistently.

Databox's AI tools (even during the beta period) have drastically changed how I manage the business. I'm able to get insight into performance into anything at any time. I can do a deep dive analysis once per week that helps me know where to focus and what questions to ask the team. It's truly like having a full-time analyst available to me, one that can answer questions instantly and never gets overwhelmed with requests.

Excited to hear how others are using this and how it'll help them.

I have been close to this product for a long time. What still gets me is watching someone use it for the first time - they come in skeptical, type one question, and within 30 seconds they are already asking a follow-up. That moment where skepticism turns into genuine curiosity is the best signal a product can give you. Really proud of what the team shipped.

Big day for the team! Really proud of what we shipped. If you are in sales and rely on numbers to hit your targets, give Genie a try - you will not go back to the old way.

Can we be trained on a specific template? For example, let's say the CMO needs the analysis done in a specific style, and templates, can we have them pre-saved?

Congrats to the whole team on this launch. It has been a long road and seeing it out in the world is genuinely satisfying.
If you are a PM tired of waiting on the data team for every small metric, this one is for you.

This is huge for my weekly and monthly marketing reporting. Dashboards are great, but Genie is faster and easier to get performance answers on the spot. I've already been using it for several weeks while in beta, and I now have a standard prompt I use to get weekly insights on our marketing leads, pipeline, and sources in less than 2 mins. Saves a tone of time AND helps me answer ad hoc leadership questions re: marketing performance much faster. Huge win!

Feels solid but a bit generic AI-powered analytics + answers fast could describe 50 tools on PH. I’d make it sharper with a specific use case or user to stand out instantly.

Top! As someone on the sales floor every day, having instant access to performance data without needing a BI tool or writing SQL is huge. I asked Genie which deals are most at risk this month and got a prioritized list right away. No report to build, no analyst to ping. Highly recommend for sales teams.

Looks pretty slick! Does it handle role-based access control across teams and dashboards? Congrats!!!

Congrats on the launch!! Is there a page I can see the integrations possible? Thanks.

Hi @davorin, congrats on the launch. You have 20,000+ customers and a wall of awards from G2, Capterra, Gartner. That's serious social proof. But it's all the way at the bottom. A founder landing on your page has to scroll past features, integrations, and use cases before they see why they should trust you.

That proof could do a lot more work for you if it hit people sooner.

Also noticed the "Unlimited users on every plan" line is buried in the "Before Databox / After Databox" section. That's a massive differentiator. Most BI tools charge per seat. That alone could be a hero line.

Just a thought. Excited to see where Genie goes. Good luck with the launch.

Congrats Databox team and Davorin. Looks great, super intuitive and can see exactly how I would use it to get immediate value. What integrations will you be supporting?

Such a cool product!

I always like being able to track back where the data source is being referenced from when AI completes their analysis. How easy is it to do that using Genie? Does it automatically always include an audit trail so that I can see where it pulled data or when it combined/manipulated data to get to a conclusion?

Congrats on the launch! Having an AI analyst that can just tell you why a metric has dropped without having to dig through a bunch of dashboards is such a huge time saver for any startup team. Can Genie pull from multiple data sources or is it limited to what's already in Databox?

The Genie AI Analyst is a smart move – asking questions in plain language instead of building custom dashboards lowers the barrier massively, especially for non-technical team members who usually depend on analysts for every ad-hoc question.

Curious about one thing: with 130+ integrations, how do you handle data consistency when different sources define the same metric differently (e.g. "revenue" in Stripe vs. QuickBooks vs. HubSpot)? Is that something Genie can flag, or is it on the user to standardize via Datasets first?

this will be a homerun application. Could use this on a small scale business to i supose?

Grats on launching. How does the tool prevent hallucinations when generating insights from live data? And for live data does the dashboards update in real time?

GPT-5.4 mini & nano are @OpenAI’s newest fast, efficient models built for real-world AI workloads.

They solve a big problem: latency and cost in production AI. Instead of relying only on large models, you now get smaller models that are faster, cheaper, and still highly capable.

• GPT-5.4 mini is 2x faster than GPT-5 mini and excels at coding, subagents, multimodal tasks, and computer use. It even approaches GPT-5.4 performance on benchmarks like SWE-Bench and OSWorld.

• GPT-5.4 nano is the most lightweight and cost-efficient option, perfect for classification, data extraction, ranking, and simple coding tasks.

What makes this powerful is the system design: use GPT-5.4 for planning, and mini/nano as fast subagents to execute tasks at scale.

With features like tool use, web search, computer interaction, and massive context, this unlocks truly responsive AI systems.

If you're building AI products, coding tools, or automation workflows, this is a big upgrade. You can use GPT-5.4 mini and nano for most of your @OpenClaw use cases.

I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

I like this direction a lot. Smaller models that still perform well feels more practical than always relying on huge ones.

Lately I've found myself switching to the 5.4 models in Cursor more often than the built-in 4.6 models. Code results are about the same, but GPT models seem to do a better job of explaining the reasoning.

Hey everyone!

I’m Prabhat Sharma, and we’re incredibly excited to share OpenObserve with the Product Hunt community today.

Observability has become a massive operational burden for engineering teams. Between managing complex, resource-heavy stacks and the technical debt of fragmented data, the status quo is broken. We built OpenObserve because we believe you shouldn't have to choose between deep visibility and system simplicity.

OpenObserve is a cloud-native observability platform built in Rust that handles logs, metrics, and traces in a single, high-performance engine. By using a stateless architecture and leveraging your preferred object storage, we’ve made it possible to manage petabytes of data with a significantly smaller infrastructure footprint than traditional tools.

Why OpenObserve?

• Built for Performance – Written in Rust with a stateless architecture. It’s incredibly fast and remarkably light on system resources.

• Storage Efficiency – Store your data on S3/MinIO/GCS instead of relying on complex block storage. Capture all your data without the need for aggressive "sampling."

• Unified Experience – Query logs, metrics, and traces using familiar SQL. No more context-switching between three different tools to find one root cause.

• Drop-in Compatibility – Use your existing collectors. We are API-compatible with Prometheus, FluentBit, OpenTelemetry, and Vector.

• AI-Powered Insights – Use our native AI Assistant to query your data using natural language or let the SRE Agent correlate incidents automatically.

Whether you’re a startup looking for visibility from day one or an enterprise seeking a streamlined "log-and-analyze" workflow, OpenObserve is built for you.

We’re in active development and would love to hear your feedback, feature requests, and critiques!

Get started:

• Try OpenObserve Cloud: https://openobserve.ai/

• Explore the Docs: https://openobserve.ai/docs/

Check us out on GitHub:https://github.com/openobserve/openobserve

Hey everyone. We built OpenObserve to fix the headaches of complex observability, but the community is what actually shaped it. The GitHub stars, the feedback, the bug reports, the honest conversations.. we owe our continued success to you.

I'll be hanging out here - happy to answer any questions!

I've been using OpenObserve for the last two years, and must say quite impressed with the features and functionality they were able to add over this time. In the beginning I longed for permissions, patterns, and parity. Lo and behold, they've implemented all that and more. Things have definitely gotten quite nice, and continue to improve daily.

OpenObserve comes through in the clutch, its alerts, notifications, and dashboards are powerful and very flexible. Ingestion is also very easy, you can either pass yourself via endpoints/kinsesis/curl, etc, or leverage standard system agents.

They are also very quick to take feedback and provide support as needed. I look forward to the new features and updates they're adding. Hopefully they remain focused on what is most needed in the platform to ensure continue growth, stability, and use.

At this point I would definitely consider it an almost drop-in replacement for most of the core DataDog logging and observability features, and in some cases they exceed them, such as the current alerts and dashboards.

There are some things I wish they supported sooner, and some things I push for them to support, but all in all a very positive experience and happy I was able to start fresh in several projects with O2 and leave DataDog (its obscene billing) and the others firmly behind.

Hi everyone! 🙋 It's been a wild journey moving from our first lines of Rust code to this launch. OpenObserve was really shaped by our early open-source contributors who pushed us to make it API-compatible with things like Prometheus and OpenTelemetry.

We’re just getting started with our AI-native features, and I’d love to hear from this community: how can we make your on-call rotations or debugging sessions less painful? Drop your questions below!

Expanding on our AI-Powered Insights: We’re big believers that AI should be a co-pilot, not a black box. While our automated SRE agent handles the heavy lifting of the analysis, we make sure to present every correlated signal. It’s all about keeping the human in the loop—giving SREs the full context they need to make fast, informed decisions with total confidence.

Congratulations

Tried a few options before this. OpenObserve was the easiest to get running and the least annoying to maintain. Logs, metrics, and traces all in one place and storage costs are surprisingly low.

For teams adopting OpenObserve today, what are some practical entry points into AIOps, are you focusing more on anomaly detection, intelligent alerting, or incident summarization first?”

Curious to know.

Didn't know the project, but going to test soon! Very cool. One question would be if there is already implementation of (or plans to add) any LLM to automatically analyze errors and report them back?

140x lower storage cost is a bold claim 👀 If this holds up in real world usage it’s a massive win for teams dealing with observability at scale.

This looks like a solid alternative to the big players! I’ve been looking for an observability tool that doesn't feel like a second job to manage. The 2-minute setup claim is bold—I’m going to spin up a cloud instance today and see how it handles my logs. Excited to see where the AI SRE agent goes!

The value prop is strong but observability tools live or die by reliability and ecosystem support observability tools live . Would love to see more benchmarks and production case studies.

Solid product! Great team and super responsive customer support - we’re a happy customer

For someone new to observability, what’s the best way to get started with OpenObserve? Is there a recommended order of operations when it comes to getting set up?

The hardest observability problem in production prediction systems isn't spotting crashes — it's detecting silent degradation. We run football prediction models where a database ingestion delay of 20 minutes means the model silently operates on stale lineups. No error thrown, no alert fired, just subtly wrong predictions for a match already in play.

The 'start in 2 minutes' framing is what actually matters here. Datadog's onboarding overhead meant most data teams gave up on proper ML pipeline instrumentation and just watched error rates. Lower friction means teams actually instrument the signals that matter — feature freshness timestamps, transform latency per stage, distribution drift — rather than just infrastructure uptime.

Curious how the AI-native angle extends to anomaly detection on custom metric streams. Is the SRE agent mostly log correlation, or does it handle time-series anomalies on arbitrary business metrics too?

Hey Product Hunt 👋,

I’m Davit, co-founder of Banyan AI, and we’re excited to launch here for the first time.

Did you know that a 5% monthly churn rate can reduce your annual revenue by nearly half? Or that many SaaS companies lose 2–5% of their revenue to leakage? New leads matter, but your existing customers are your real treasure.

If you're struggling with churn or finding it hard to expand revenue, we’ve got you covered. Welcome to Banyan AI 🌳🚀

Our platform unifies data across your tool stack (billing, CRM, product analytics, support) and detects signals that are scattered across those tools. Banyan AI automatically detects:

• Customers likely to churn

• Hidden revenue leaks

• Expansion opportunities

Instead of digging through dashboards, you get clear AI insights about your revenue health and what needs attention. Check out our website or our blog.

I’ll be here in the comments all day. Thanks for checking out Banyan AI 🙏

In education, churn looks different than in typical SaaS — a student finishes a course and leaves, which is not churn, it is a natural end. But someone who stops halfway through is a completely different signal. How well does the detection handle that difference, where inactivity does not always mean risk?

nice one honestly. most teams talk a lot about getting new customers, but keeping the existing ones is where so much revenue gets lost quietly. i like that this seems focused on helping teams spot the risk earlier instead of just reporting on it later.

curious, are people finding more value in the churn detection side first or the expansion opportunity side?

this is a useful space to be building in. a lot of saas teams only realize churn is becoming a real problem once the damage is already visible, so bringing billing, product, crm and support signals together makes a lot of sense.

curious, what kind of signal ends up being the strongest early warning most of the time, product usage drop, support issues, or something else?

Good luck team! Cool idea. How long does the setup usually take if someone wants to connect their SaaS tools?

Clear and compelling spotting churn before it happens is an immediate pain point for SaaS teams. I’d punch it up with the main outcome first: “Prevent churn and capture expansion opportunities in minutes, all from one unified dashboard.

Strong problem, but the hook feels a bit textbook SaaS I’d cut the stats and lead with a sharper outcome like spot churn 30 days before it happens. Time to value: minutes is great though lean harder into that and may be show one concrete example to make it feel real.

Great product that solves a real pain! Reducing churn is low-hanging fruit as long as you know how to detect it beforehand :)

Looks awesome. Congrats on the launch!

All the best guys, thats really a pretty cool idea!

Congratulations on the launch 🎉 🎉

The "text-to-API" approach for connecting data sources is clever – removing the integration bottleneck is probably the single biggest thing that determines whether a tool like this actually gets adopted or sits unused after the trial.

One question: at what point does Banyan become useful in terms of data volume? If I'm an early-stage SaaS with 50-100 customers, is there enough signal for meaningful churn predictions, or does this really shine once you hit a certain scale?

What signals do you look for to identify customers that are likely to churn?

As one who is building my first SaaS product, this is really interesting! Particularly intriguing is the ability to identify clients or customers who are ready to upgrade.

Well after checking out this product and seeing the potential churn behind our own, this one helped us save a decent amount of our churn by being able to assist our own customers.

Unifying billing, CRM, support and product usage into one view is the real value here. Most founders check these in 4 different dashboards. Congrats!

Being proactive as oppose to reactive is definitely the way to go. I'd love to see progress into remediating and taking actions to prevent churn. Congrats on the launch!

Hey, congrats on the launch! Churn is such an underrated problem in SaaS, and I like how you have combined all this data into one view. Makes a lot of sense. Curious to try it out! Do you get like instant notifications?

Congrats on the launch, @davitausberlin.

That 5% churn stat you shared really hits me. I spent some time on your site. The comparison table with Gong, ChurnZero, and Clari caught my eye. That's bold. Most companies hide from comparisons. You put it right out there. I respect that.

One small thing I noticed. That table is tucked under a Comparison tab. A founder skimming the page might never click there. If it's one of your best trust builders, maybe give it more space.

Also the "2 Minutes from Sign-Up to Churn Detection" line is gold. That's exactly what busy founders want to hear.

Good luck with the launch. Excited to see where this goes.

Sounds interesting! How does it alert you about who is about to churn — through email, in-app notifications, or something else?

I've watched Davit and Konstantin work on Banyan over the past few months, and I have to say, what they're building is incredible. A business isn't sustainable if it can't tackle churn. Banyan makes it easy to tackle it head-on in an easy and accessible way. Having everything you need all in one place makes that possible, let alone the clever use of AI agents to make life as easy as possible! Looking forward to following you both and how Banyan evolves in to the future, Davit and Konstantin. This is the start of something special, i'm feeling it!

Hey Product Hunt 👋,

I’m Konstantin, CTO and co-founder of Banyan AI.

If you want to detect customers who are about to churn, or identify those with the highest expansion potential and real impact on your revenue, forget about building your own dashboards, stitching together dozens of APIs, or copying data into Excel.

We built Banyan AI so you don’t have to do all of that. With our tool, everything works within minutes — fast time to value, no technical skills needed. You ask, Banyan answers.

It sounds simple, but there’s a lot of engineering behind the scenes.

Happy to answer any technical questions here 👇

congrats on the launch!

How many customers must you have before you start valuing a tool like this? Congrats on the launch!

Interesting positioning. Detection is everywhere right now, but prevention is where most tools fail. Curious what actually triggers action in your system, not just insights.

Congratulations on your launch!

How flexible is the template system when controlled by an AI agent? Can it handle dynamic personalization easily?

this is actually a pretty smart direction. a lot of ai workflows sound great until email comes into the picture and suddenly you need extra setup, delivery infra, and a bunch of manual steps. making all of that accessible inside the same workflow feels really practical.

curious, are most people using this first for transactional stuff or proper outbound/campaign use cases?

Are people mainly using this for email campaigns , or does it extend to other channels as well ?

This is cool. We use autosend a lot for both transactional & campaign email.

How much human review is typically needed before sending out a campaign?

Cool concept letting your AI handle emails end-to-end is a neat shortcut. Could land stronger by leading with the benefit: Let your AI run campaigns, manage audiences, and track results automatically.

Congrats on the launch! Email infrastructure inside your AI workflow is a game changer. Does the MCP sever support multi-step sequences or just one-off sends?

Super happy to see this finally live! Agentic experience is a must have right now, and having it added to the emails is extremely useful. Proud of the team 🥳

Nice idea. Embedding email sending directly into AI workflows without extra setup feels like a natural step for agent-based apps. The MCP integration across different tools is a strong touch. Can AutoSend automatically personalize transactional emails based on user data passed by the agent, or does that still need manual setup?

The idea is powerful but email marketing is sensitive. one mistake can hurt deliverability would be great to see guardrails and approval workflows?

This is a strong problem to go after. A lot of teams are excited about MCP, but the security and authorization layer is exactly where things start getting uncomfortable once real production access is involved. The fact that this works as a proxy and does not require rewriting agents or servers makes it feel much more realistic for actual adoption.

Curious, what tends to be the biggest blocker for teams right now when they start thinking about MCP security, visibility, fine-grained control, or integration with existing identity systems?

Hey PH ! Or Weis here, co-founder and CEO of Permit.io. Fourth time launching here, and always great to be back.

We’ve been building in authorization for years, and the shift we’re seeing with MCP feels like one of those rare infrastructure moments. Every protocol starts a little messy. HTTP was messy. TCP/IP was messy. MCP is no exception. But it is quickly becoming the connective tissue between AI agents and enterprise systems, which makes it the right place to enforce identity, trust, and governance.

Most of the market looks at MCP and asks, “How do I push this through my existing stack?” We think that is the wrong question.

Agents are not service accounts with better branding. They need a new kind of identity: dynamic, delegated, auditable, and revocable in real time.

That is why we built Permit MCP Gateway.

Permit MCP Gateway is a drop-in trust layer for MCP. It helps teams secure AI agents connecting to tools and enterprise systems with fine-grained authorization, consent, auditability, and runtime enforcement — without rewriting their stack.

A few things we think matter:

• fine-grained permissions for agent actions

• delegated access on behalf of users

• audit logs for every tool call

• zero-standing-privilege approach

• built on Permit, so controls can extend deeper into APIs, services, and data for defense in depth

This is a very natural evolution for us. Permit started with application authorization, and now we’re bringing the same philosophy into the AI era.

If you’re thinking about how to bring MCP into your organization without turning your systems into open desert, we’d love to talk.

We’re here all day — would love your feedback, questions, and skepticism.

Hey Product Hunt! David here, Solutions Engineer at Permit.io.

We just published two walkthroughs showing the MCP Gateway in action:

Enforce per-user trust levels on Linear's MCP (Developer vs PM access): https://docs.permit.io/permit-mcp-gateway/demos/linear-mcp-gateway

Gate an n8n automation workflow with real-time trust controls: https://docs.permit.io/permit-mcp-gateway/demos/n8n-linear-mcp-gateway

No changes to the underlying MCP servers — just drop the Gateway in front and control who (or what) can do what. Both demos take just a few minutes to set up. Would love to hear what MCPs you'd want to see demoed next!

okay yeah this makes a lot of sense. everyone wants agents to connect to tools now, but the second you think about who approved what and what that agent is actually allowed to do, it gets serious real fast. the one url change part is probably what will make people actually try it.

curious, what’s the first reaction you get from security teams when they see this, relief or more questions?

I like the direction here. Instead of building new systems, you’re improving what already exists and making it safer.

The consent and authorization flow sounds strong. I’d just want to make sure it stays simple for end users and doesn't add too much friction.

How easy is it to roll this out across multiple MCP servers in a larger setup? Especially for teams managing different environments.

What stands out to me is the “no code changes” part. A lot of security tools require heavy integration, so just swapping a URL and getting full authorization control sounds very appealing.

It would be helpful to see a simple dashboard or visual layer where teams can quickly understand who has access to what without digging into logs.

One thing I’d want to understand better is performance overhead. Since every request goes through the gateway, does it introduce noticeable latency in high frequency systems?

Wow, this is really impressive 👋I love that you’re tackling such a tricky but crucial part of the stack , authorization and security are easy to overlook until something breaks. The fact that it works as a transparent proxy without touching agents or servers makes it feel much more approachable for real world use.

I’m curious too , how do teams usually start tackling the visibility and fine-grained control challenges? It seems like understanding what agents can do, and who’s authorizing them, is where most of the headaches come in.

Hmm , MCP is powerful, but auth and governance are definitely the messy part right now. Adding OAuth and fine grained permissions without changing the server code sounds super practical. The “just swap the URL” part is especially cool. Congrats on the launch .

This feels like a must have for teams building AI agents or MCP based tools .Sharing this with couple of backend/security folks .

How does this handle performance under heavy load ?Zero latency sounds great curious what that looks like in real world benchmarks .

This is Great!

Def gonna try this out!

The concept is strong but security tools live and de by trust .Would be great to see audits compliance certifications or deeper technical docs.

Having audit trails is so important, so having the ability to know who authorized which agent is really nifty. Does Permit.io flag when policies fall outside standard best practices? Or does the auto-generation capability fully manage this such that no manual configuration is required after set up?

Agent interrogation - seems interesting but problematic, how can you trust the agent not to lie, or be coerced to lie ? How can this produce a consistent Identity?

Hi everyone!

Unsloth Studio just made fine-tuning models way more intuitive and accessible.

Instead of writing complex training scripts, you now get a clean GUI for the full workflow — dataset management from PDFs/CSVs, training config, real-time monitoring, and even auto dataset recipes. Everything stays local and private.

Unsloth already had a HUGE following for its super-efficient LoRA fine-tuning. Studio turns that into a full platform, lowering the barrier so way more people can experiment, customize, and have fun with their own models.

This is going to accelerate the whole open model scene!

It might be useful to have a few ready-made voice profiles for quick testing. Helps a lot when you just want to plug and play.

I’ve seen tools that handle parts of this workflow, but having everything combined in one UI feels more practical than jumping between setups.

Fine-tuning models feels way faster now. Less memory use makes testing multiple setups so much smoother.

Nice local AI plus and no code training hits a growing niche for privacy-conscious devs. I’d lead with the main benefit: Train and run AI models locally, faster and with less VRAM, no coding required.

great job guys, cant wait to try this

ngl the best part here is probably that it keeps things local. people are way more interested in experimenting with their own data now, but they still want privacy and don’t always wanna fight with setup for hours just to test something.

also curious, what kind of files are people usually starting with first, pdfs, docs, support chats or something else?

I know a couple of people who work with local models and are always looking for ways to simplify training. This looks like something I'd share with them.

I’ve tried fine-tuning models before and honestly gave up halfway because of the setup. This feels like something that would actually make me try again.

I’m curious how it handles larger datasets or longer training runs. Does performance stay stable as things scale?

How well does it handle large date sets and longer training runs locally ?Any limits users should be aware of?

Tried Unsloth to fine-tune some Qwen model with LoRA, loved it :)

This is super cool. have been thinking of building something like this for long time.

This could be a really interesting way to help companies better customize their own LLMs based upon their own company's context, processes, and culture. Is use of the Unsloth's Data Recipes the best way to tackle something like this?

this feels like a pretty important step for open models honestly. a lot of people want to fine tune their own models, but the moment they see scripts, configs and setup headaches, they drop the idea. making the whole flow visual and local could open this up to way more people.

curious, who are you seeing get the most value from studio so far, hobbyists learning this stuff or teams actually training models for work?

Excited to check it out - this feels so much more accessible to us folks who are semi-technical

Love the idea but 2x faster +80% less memory sounds almost too good would love to see benchmarks across different models.

Overall, this looks like a solid addition for developers building voice features. If the controls are easy to use and consistent, I can see this being widely adopted.

Hello, this looks promising. The expressive controls part really stand out because voice quality alone isn’t enough if tone and delivery feel flat.

Natural voices make a big difference, especially for anything users listen to regularly.

It would be great to have some preset voice styles (like conversational, formal, energetic) for quick use instead of configuring everything manually.

I’ve used a few TTS tools before, and even when the voice sounds good, the lack of control over pricing and emotion becomes noticeable quickly.

Getting the same identity across different styles in not easy. This looks solid.

Curious what kind of use cases you’re seeing the most so far. Is it more around voice agents, or things like content narration and apps?

this looks useful. the expressive controls part is what caught my eye because that usually matters as much as the voice itself.

curious, are you seeing more use for voice agents or general product experiences?

This is excellent news! The promise of natural voices and expressive controls is particularly appealing. I'm curious about the range of languages and accents supported, and whether there are options for custom voice branding.

The pricing page shows $0.10 per 1000 characters - that's roughly 2-3 cents per typical API response. For a side project handling 10k requests/month, we're talking coffee money, not rent money. Refreshingly honest pricing in a space where everyone else hides behind "contact sales.

This looks cool, can we clone the voice with custom premium voice with cinematic background and some sort of frequency settings?

Only English is supported for now, right?

nice one. natural sounding voices make a huge difference when people are actually going to listen for more than a few seconds.

curious, what kind of apps are people building with it first?

Soul 2.0 is a game-changer for anyone tired of plastic-looking AI photos.

The magic is in Soul ID. You train it once with your photos, and you get consistent character fidelity across wildly different styles, lighting, and poses. That's not trivial.

50+ presets mean you skip the entire "prompt engineering hell" that kills most creators. Pick a vibe, write naturally, and boom. Magazine spread quality.

Who's this for? UGC creators building ad funnels, fashion brands prototyping lookbooks, indie artists moodboarding without a crew. Basically anyone who needs photorealistic consistency at scale.

The thing that moves the needle: it actually feels shot, not generated. Real grain, real light falloff, real fabric texture.

This is honestly great for someone like me who loves fashion but hates the logistics of photoshoots. It lets me try hundreds of looks in minutes and everything feels real and polished. I’m exploring styles I would’ve never dared before.

this looks pretty solid. the consistency part is what really stands out because that’s where a lot of ai photo tools still fall apart. if soul id can actually keep the same person feeling real across different looks and setups, that’s a big deal.

curious, what are people using it for most right now, creator content or proper brand shoots?

I’ve tried generating images for content before, and keeping the same face consistent across output is always a struggle. This seems to solve that directly.

Hello, this looks really impressive. The consistency part is what stands out to me because that’s usually where most AI photo tools fall short.

Curious how flexible the presets are. Can users tweak them slightly, or are they more fixed styles?

I know someone working on ad creatives who spends a lot on shoots just to test variations. This feels like something they'd want to try.

The consistency challenge is huge for branding. One great AI image is easy but keeping the same character, lighting, and vibe across a whole campaign is tough. If Soul ID can do that, it’s a real game changer.

Solving the consistency problem is genuinely interesting from a brand perspective. I work in branding and the single biggest friction with AI imagery for brand use is that you can generate one great image but can't 100% reliably reproduce the same character, lighting feel, or aesthetic across a campaign. If Soul ID holds across wildly different setups that's a real workflow unlock for brand shoots, not just creator content.

Curious about the edge cases: how does Soul ID handle significant changes in lighting direction, like moving from a softbox studio setup to harsh outdoor midday light? That's where identity lock tends to break down and where brand shoots get complicated.

nice one. the part about it feeling shot and not generated is probably the most important thing here. a lot of ai images still have that weird artificial look, so getting the lighting, texture and identity consistency right changes a lot.

curious, which part took the most work to get right, the identity lock or the final photo realism?

The output looks impressive but with AI generated visuals consistency over time is key .Would love to see term usage example.

This is genius. The character fidelity is the biggest challenge in AI images and video. As much as I value the traditional production process, imagine all of the small brands and creators yet to be who can't afford production costs. We need to spread the narrative of AI focused on empowerment, entrepreneurship, and what it can do for people's dreams as opposed to what it takes away!

Comet Enterprise is an AI-powered browser built for teams, combining research, automation, and workflow execution in one place.

It solves the chaos of tab overload and fragmented tools by bringing context-aware assistance directly into your browser so teams can research, summarize, automate tasks, and prepare work without switching apps.

What makes it stand out is its enterprise-grade foundation: granular admin controls, telemetry, audit logs, MDM deployment, and deep security integration with CrowdStrike’s Falcon platform to protect against phishing and malware.

Key features:

• Context-aware AI across tabs

• Task automation (emails, meetings, workflows)

• Centralized deployment & visibility

• Advanced security + compliance (SOC 2, HIPAA)

• Admin-level control over access and usage

Benefits:

• Save time on repetitive work

• Reduce tool-switching

• Improve team productivity securely at scale

Best for: Enterprise teams, ops, analysts, and organizations managing complex workflows across multiple systems.

Strong step toward the “AI-native browser” future for teams. Do you think AI browsers will replace traditional enterprise workflows soon?

I hunt the latest and greatest launches in tech, SaaS and AI, follow to be notified → @rohanrecommends

Solid positioning Where Knowledge Begins is clean and aspirational. Could be punchier by leading with the key benefit: Get answers and research insights instantly, wherever you are.

this actually makes a lot of sense tbh. teams already have too many tabs, too many tools, too many places where work gets split up. doing research + actions + workflow stuff inside the browser itself feels way cleaner than bouncing around everywhere.

curious, what’s the first use case people usually get hooked on with this?```

This is an interesting direction. A lot of enterprise teams are already living inside the browser all day, so combining research, automation, and execution in one secure workspace feels pretty natural. The security and admin layer is probably what makes this much more viable for real team adoption.

Curious, what kind of workflow are teams finding most valuable first, research-heavy work or more operational task automation?

Finally, a calendar app that understands "coffee with Sarah next Tuesday at that place we went last time" without requiring a PhD in click-through navigation. The natural language parsing is surprisingly tolerant of my 2am typo habits - it correctly interpreted "dentst appt tomorrw 3pm" which is either impressive or deeply concerning about how often I reschedule dental appointments.

As a years-long Fantastical subscriber, I am super stoked about this. I can't wait to play!

okay this is actually cool. calendar stuff always looks easy from outside, then one vague request or timezone mixup turns into a mess real fast. getting this to work naturally with claude must have taken way more effort than people think.

curious, what kind of command do people try first when they test it? recurring events, reminders, or just basic scheduling?

This is a really nice fit. Scheduling is one of those things that sounds simple until you get into time zones, recurring events, and vague requests, so getting it to feel natural inside Claude is genuinely impressive. It also helps that Fantastical already has such a strong reputation for good calendar UX.

Curious, which edge case ended up being the hardest to make feel reliable in real usage?

I've been waiting for the Fantastical MCP! This is great!

Would be interesting to see support for sources like local files or browser history. That could make recall even more powerful.

It could be useful to have optional lightweight syncing, maybe encrypted, for people who work across multiple devices.

I’ve noticed I hesitate before writing certain things in cloud-based tools. That small pause you mentioned is real, and it does affect how freely you think.

Overall, this feels like a thoughtful direction. If the local setup stays simple and performance holds up, I can see this becoming part of daily workflows.

Nice launch. I like the system tray approach — it feels less intrusive than opening a full app every time you want to jot something down.

Using Ollama LanceDB for offline RAG is a solid stack. Nice to see thoughtful architecture behind the scene.

This makes a lot of sense. Most of our computers already function as a personal knowledge base, even if it's completely disorganized. It will take some time for trust to warm up to tools like this, but they seem inevitable. Even if the AI runs locally, having a way to easily wipe the AI's memory (clear cache, like a web browser) should provide peace of mind.

Love the idea of a private, local “second brain” feels like the direction personal AI should be heading.

Simple, fast, and no tracking is a big win.

Curious, how does Lore handle context over time as data grows?

Great work 👏

Love seeing more open-source approaches to personal knowledge management! The privacy angle is huge right now, especially with AI reading our notes. How are you handling syncing across devices while keeping everything 100% private?

ngl the "privacy tax" line hits. a lot of note and memory tools sound cool till you realize you’re basically handing over your whole brain to some cloud service. keeping it free, local and open-source makes this feel way more honest.

also curious, what made you choose the system tray quick capture style instead of going for a bigger full app first?

This is a thoughtful direction. The idea of having an AI that understands your own notes and history is powerful, but the privacy concern is exactly what stops a lot of people from fully trusting these tools. Keeping everything local makes the whole product feel much more convincing.

Curious, what kind of users are connecting with Lore the fastest so far, people using it for work knowledge or more personal thinking and journaling?

Does the folder structure stay intact if Google updates NotebookLM or is there a risk of things breaking with UI changes?

Congrats on launching and building! 🔥

This resonated with me because I'm also at a point where I have a lot of things in my NotebookLM, and I'm trying to find a way to keep everything organized 😅

This is genuinely useful. NotebookLM becomes a lot harder to manage once the number of notebooks starts growing, so adding folders, search, and better organization feels like a very natural extension.

Curious, was folders the biggest pain point from users, or did search and sync come up just as often?

Hi everyone! I'm the maker of Bookshelf 👋

I built this because managing sources in NotebookLM quickly became messy as I used it more.

Bookshelf adds folders, search, and structure — making research workflows much easier.

Fun fact: I'm actually not a professional engineer. I built this using AI tools and a lot of trial and error.

Would love to hear your feedback or answer any questions 🙌

Hey Product Hunt! 👋

I'm Vivek, the maker of ClawMetry. A few weeks ago, I launched the open-source version here and you gave an incredible #5 Product of the Day. Thank you!

But after talking to dozens of users, one thing kept coming up: "I love the dashboard, but I can't see it when I'm away from my desk."

That's the problem ClawMetry Cloud solves. It syncs your AI agent metrics to the cloud so you can check on your agents from any browser, anywhere. The catch? We built it with end-to-end encryption, your data is encrypted before it ever leaves your machine. Even we can't read it.

What's new in Cloud:

• 🔒 E2E encrypted sync (zero-knowledge architecture)
• 🌍 Browser access from anywhere (https://app.clawmetry.com/)
• 📊 Multi-node monitoring (all your agents, one dashboard)
• 🍎 Native Mac app

Some context on traction:

• 75K+ PyPI downloads
• Built for the OpenClaw ecosystem (316K GitHub stars)
• 7-day free trial, then $5/node/month

I'd love to hear: What's the biggest pain point you have monitoring your AI agents in production?

Happy to answer any questions! 🙏

okay this is cool. “can’t see it when i’m away from my desk” is such a real product moment lol. a lot of tools feel fine till you actually need to check something quickly from outside, then that gap becomes super obvious. also love that you didn’t ignore the privacy side while adding cloud sync.

curious, what’s the first thing people usually open clawmetry cloud to check when they’re away?

Solid tool real-time observability for AI agents is super practical, especially for devs managing multiple sub-agents.
Could land stronger with the key benefit up front Monitor your Open Claw agents live costs, activity, and memory without any setup.

This is a smart extension of the original product. Local observability makes sense at first, but the moment agents start doing real work, being able to check on them remotely becomes much more important. The end-to-end encryption angle makes this a lot more convincing too, because monitoring data can get sensitive very quickly.

Curious, what tends to matter most to users once they go remote, uptime visibility, cost tracking, or understanding agent behavior over time?

congrats on the launch!

Having E2E encryption built right into the monitoring layer is a massive relief for anyone running OpenClaw agents on untrusted networks. I would definitely use this to keep tabs on my remote scraping nodes without having to tunnel into each server manually. I am really curious to hear how you handle the initial key exchange process when provisioning a new agent from the command line.

Hey Product Hunters 👋, I'm Berat.

I've been building and publishing apps for over 8 years. Developing apps is the fun part, but dealing with the slow web interfaces that come after the build?

Managing releases, writing metadata for 40 localizations, creating screenshots, setting up legal pages... it's a tedious, multi-tab nightmare that eats hours you could spend coding.

That's why I brought Forvibe to your dock a native macOS app designed to handle everything between your Xcode build and the App Store. I built this specifically for Mac because your launch workflow shouldn't be interrupted by browser lag or clunky web portals.

With Forvibe for macOS, you can manage your entire release lifecycle without ever opening the App Store Connect website:

• Native Speed: Edit app metadata and localize it across all supported languages in one click, right from your desktop.

• Built-in ASO Engine: Discover high-performing keywords and optimize your visibility directly within the app.

• Store Assets & Pricing: Set up In-App Purchases with smart, country-specific pricing, design stunning screenshots with ready-made templates, and push them directly to the store.

• Instant Web Presence: Generate a professional landing page in seconds (no domain, no hosting, no code) and instantly publish AI-created legal pages like Privacy Policy, Terms of Use, and EULA.

• Unified Support: Manage customer feedback and app reviews with our AI-powered native inbox, helping you respond faster without context switching.

In short: build your app in Xcode, and let Forvibe for macOS seamlessly ship everything else.

I’d love to hear your feedback and answer any questions you have! 🚀

Love the idea but I'd be curious how customizable everything is.Automation is great, but sometimes store listings need very specific tweaks to perform well.

So basically this removes the part of app development we all procrastinate on the 😅

This would 've saved me hours on my last app launch. Especially the screenshot and pricing parts those always take way longer than expected.

How deep does the localization go? Is just translation or does it adapt screenshots and copy based on cultural context too ?

Would love to know when a windows version is coming.

The localization and screenshot generation piece is what gets overlooked most. You spend weeks on the app and then rush the store listing in a day because you're exhausted. Anything that removes that friction is worth it for solo devs. Does it handle App Store Connect metadata versioning or is it one-way push?

The PRODUCTHUNT60 code doesn't work.

okay this one hits. shipping apps is fun till you get stuck doing all the annoying app store stuff in 20 tabs for hours. having one native mac app handle that mess sounds super useful.

curious, what’s the feature people get hooked on first when they try it, screenshots, pricing, or the landing page part?

This is a very real pain point. Building the app is usually the exciting part, but everything that comes after that like store listings, localization, screenshots, pricing, and legal pages can easily turn into a full workflow on its own. Bringing all of that into a native Mac app makes the whole pitch feel much more compelling.

Curious, which part saves developers the most time in practice, localization, screenshots, or release management itself?

this hits a real pain tbh. most people don’t wanna open 5 tools just to answer one simple question like “are we actually making money from this?” the ads + revenue + web vitals combo is what makes this feel different to me.

curious, did you build this more for saas first or ecommerce first?

This is a smart angle. A lot of founders don’t really have an analytics problem, they have a fragmentation problem. Revenue is in one place, ad spend in another, product behavior somewhere else, and making sense of it all takes more effort than it should. Bringing those signals together in one view feels genuinely useful.

Curious, what ends up being the biggest aha moment for users once they connect everything, ad ROI, churn patterns, or something else?

So this is basically the finally I know what''s going on in my in my business dashboard 😅

For early stages SaaS this could replaces 3_4 tools easily .Stripe analytic error monitoring alone is a big win.

The all in one angle is great but dashboards like this can get overwhelming fast. Curious how you're handling simplicity vs depth.

Congrats on the launch! 🎉

The "stop juggling tools" angle really hits home — switching between GA4, Stripe, and ad dashboards every morning is genuinely exhausting.

One question: does MetricMap support tracking conversions from referral sources like Product Hunt or Reddit? As a maker tracking where my users actually come from, that would be the feature that gets me to switch immediately.

The web vitals + revenue combo in one view looks really powerful. Upvoted! 🚀

How accurate is the attribution between ad spend and actual revenue ?Especially across multiple channels like Meta, Google, etc.

Quick Update: New Landing Page Live! 🚀

I’ve just updated our landing page to better showcase our ecosystem. We’re doubling down on integrations to finally kill that fragmentation pain:

✅ Live: Stripe, Paddle, Lemon Squeezy, Polar, Shopify, Dodo Payments.

🛠️ Coming Soon: Meta Ads, Google Ads, TikTok, Instagram & Threads.

Now you can see exactly how your revenue connects to your tech health in one view.

Check out the new look (and grab your 20% OFF for 3 months): 👉 MetricMap.tech

Really like the direction but I'm curious what makes this stand out vs existing tools like Docusaurus or Maintlify?the Feels like a crowded space, so differentiation will be key.

Solid for devs turning Markdown/MDX into AI-ready docs with collaboration is handy. Could land stronger with the key outcome up front build beautiful AI powered docs from your files, with team collaboration and custom domains.

Hi Product Hunt! I'm Luan, one of the founders here at Doccupine.

I've always struggled to find solid documentation tooling. Most solutions are too expensive or too complex. If that's not the case, they're locking you in with proprietary tech. None of them let you bring your own AI model or integrate your docs directly into AI development workflows.

Doccupine is an open source CLI (npx doccupine). It turns Markdown and MDX files into a complete documentation website. You can self host it on your server for free. You can also bring your own AI model. To start, we have 12+ popular models from OpenAI, Anthropic, and Google (and we will be adding more). Plug in your API keys and go. Doccupine also comes with an MCP server out of the box.

We're monetizing through our managed platform. It's for teams who want user roles and permissions, easy front-end editing, managed AI with transparent budget caps (you can also still bring your own AI model), zero-config deploys, and automatic updates. We're bootstrapped. Two founders, no VC. We answer every support email. 

We'd love your feedback. What would make Doccupine useful for you?

How does it compare to @Documentation.AI?

Congrats on the launch! 🎉

The MCP support is what really caught my attention — most documentation tools treat AI as an afterthought, but baking it in from the CLI level is a smart architectural decision.

Quick question: how does Doccupine handle versioning for docs? For example, if I'm documenting a web tool that ships updates frequently, can I maintain v1 and v2 docs side by side without duplicating everything?

The Markdown → beautiful docs pipeline looks really clean.

Upvoted and following to see where this goes! 🚀

Great launch, @luangjokaj. Open source doc tool with MCP support. That's something I don't see every single day.

I spent about 10 minutes going through the site. One thing caught my attention. You're leading hard with AI features. 12 models. Bring your own. MCP server. All good things. But here's the thing. Devs are tired of AI being the main character in every product. The part that got me excited was "npx doccupine" and done. That's the dream. No setup. No config.

Also looked at your pricing. $200/month for Pro with one project. Compared to self-hosting for free, I'm not sure what I'm paying for. AI-Powered Documentation Assistant sounds nice but what does it actually do? If it's just plugging in my API key, that's a tough sell.

Curious how you're thinking about this. Either way, cool to see someone building in the open.

this is actually pretty nice tbh. docs are one of those things teams need badly, but the tooling around them often ends up being weirdly costly or too opinionated. i like that this gives people a simpler path and still lets them use their own ai setup.

curious, what kind of teams are getting into it fastest right now, solo devs shipping docs or proper teams replacing older tools?

This is a strong direction. A lot of documentation tools either get expensive fast or box teams into their own way of working, so the open-source plus bring-your-own-model angle makes this stand out. The MCP support also makes it feel more aligned with where developer workflows are heading.

Curious, are most people starting with the self-hosted CLI route first, or going straight to the managed platform?

Does it also track the normal video views or it just does it for YouTube shorts? Is Instagram on the roadmap?

Who is the target audience? The creator or the company that collaborates with creators?

I can see agencies running UGC campaigns benefiting a lot from this.Automating payouts alone could save hours every week.

Which platforms are supported right now for pulling view data? Just YouTube or also TikTok/Instagram?

This solves a very real pain. Managing creator payouts in spreadsheets gets messy fast especially with performance based campaigns.

Hey @begaiym_adylbek_kyzy, congrats on the launch. I've worked with a few agencies running UGC campaigns and spreadsheets are still the default. It's painful.

I spent some time on your site. A few things stood out to me.

I like the "save 5-10 hours per week" stat. It's your strongest hook. But it's buried under a "Why ClipLedger" section halfway down. That number should be in the first sentence someone reads.

Also, the pricing page shows $19 and $49 but doesn't explain what active tracking means. If I'm an agency with 200 videos, do I need Pro? The jump from 100 to 500 to unlimited makes sense, but active tracking could be clearer.

And lastly the MVP approach is unique. Building with early users is the right move. I'm curious how you're handling creators who submit the same video to multiple campaigns. Is that something you're thinking about?

Either way, excited to see where this goes. Good luck with the launch!

this is actually useful tbh. anything tied to creator payouts gets messy real fast once multiple people and campaign rules are involved, and spreadsheets only work till a point. i like that you kept the scope tight with youtube shorts first.

curious, are early users mostly agencies or in-house teams managing ugc creators?

This is a pretty practical problem to solve. Once creator campaigns start scaling, payout tracking turns into one of those messy backend tasks that eats way too much time and still leaves room for mistakes. Focusing on YouTube Shorts first also makes sense instead of trying to cover everything at once.

Curious, what part do teams struggle with most right now, collecting the view data or agreeing on payout rules?

Love the concept but accuracy is everything here .How do you handle delayed or fluctuating view counts across platforms?