PH热榜 | 2026-04-25

Hello PH fam!

To be honest, I originally built ZeroHuman as a tool for myself, but I wanted to share it with you all as well.

It's every founders dream to have a "lab" where you can spin out tons of products, but this was never really possible before because it's just too hard to even run one company. However, AI is now advanced enough to make this dream make this a reality.

So we built ZeroHuman to combines 4 tools to let you create your own "zero human" companies:

1) OpenClaw: Chat with your AI Co-Founder
2) Paperclip: Manage tasks for your company
3) GPT 5.5 Spud: Code with the frontier model
4) Cracked.ai: Market with automated viral reels

AI is potentially coming for everyone's job.
Our mission is to make AI work for you instead.

Hope you enjoy, let us know what you think!
Dan

Congratulations the launch. @dansutera

Two years ago, Sam Altman predicted we'd see the first one-person billion-dollar company. At the time it sounded like Silicon Valley hyperbole. Today it sounds like a roadmap. Pieter Levels is running a $3M+ ARR portfolio as a solo founder with zero employees. Midjourney hit $200M ARR with ~11 people. The bottleneck isn't ambition or capital anymore, it's bandwidth.

That's the gap Zero Human is filling:

– OpenClaw (361k+ GitHub stars) for the always-on, mobile-first chat layer – your co-founder lives on your phone

– Paperclip (57k+ stars) for orchestration – the structure that makes it feel like a real co-founder, not just a chatbot

– Cracked for the daily content engine – short-form video on TikTok, Instagram, YouTube

Tried it myself a few days ago. What stood out is the clarity. It actually generates documentation as you go (mission, PRD, daily reports), so you have a real reference point for where you've been and where you're heading. Zero Human gives you a map for the journey.

One meaningful project per day, runs at night, you wake up to a briefing. This is one of the first products that actually feels like having a co-founder, not operating a tool.

Honestly as a solo founder, dad of 3, full time job, the bit I'd care most about is whether this thing actually pushes back when I'm wrong or just hypes everything I throw at it. Last thing I need is a yes man. Does it ever properly disagree with you, call out a bad idea before you've burnt a weekend on it ?

congrats! would be awesome if ZeroHuman did LinkedIn and X posts, as well :)

I am giving up on creating a product. :D AI surpassed me in everything.

Congrats on the launch! How does the short form video thing work, didn't find much info on your website

@benln @ZeroHuman. Getting an error trying to start. Its a cors error from your backend api : 1/onboarding/sessions

The "AI co-founder" framing is interesting and I think it reflects a genuine shift in how solo builders operate. The real leverage isn't just automating tasks — it's having something that holds context across the different functions of running a product (marketing, user research, ops) so you don't have to context-switch constantly.

I've experienced this firsthand building a few side projects alongside my day job in M&A. One is DishRoll (https://dishroll.netlify.app/) — an AI-powered weekly meal planning PWA that generates personalised plans and shopping lists. The only way it got built was by treating AI as a genuine collaborator on product decisions, not just a code assistant.

Curious whether ZeroHuman maintains persistent context across sessions or whether each task is relatively standalone? That continuity piece feels like the hard problem for AI co-founder products.

Given how eager LLMs are to generate placeholder testimonials when creating landing pages, I have to ask: are the testimonials on the home page genuine?

The Cracked.ai viral reels piece is the one I'd love to understand better. Automated marketing is easy to build badly and hard to build well. How does it avoid the AI-generated look on TikTok and Reels? Does it pull from your actual product, or is it more template-driven?

Gemini now reads your Gmail, Photos, and YouTube before answering.

What it is: Personal Intelligence is a personalization feature inside Gemini that connects to your Google apps to generate responses grounded in your actual history.

The problem it solves: AI assistants are generic by default. They know everything about the world and nothing about you. Every session starts cold. Personal Intelligence changes that by making your Google activity available as context for every response.

Key features:

• Connects to Gmail, Photos, Search, YouTube

• Uses past Gemini chats for continuity across sessions

• Custom standing instructions for goals and preferences

• Per-app controls you can adjust or revoke at any time

Benefits:

• Ask for recommendations and get answers filtered by your real preferences

• Reference your own emails or photos without pasting them in manually

• Pick up previous conversations without summarizing them yourself

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

Awesome

Hey everyone, Etienne here, I'm the CEO.

I want to add some context on the commercial side because it tells part of the story Kshitij didn't cover.

When we were last on Product Hunt, we had a handful of early customers and, honestly, more conviction than proof.

Today, we're at 10,000+ active users with accounts at Loréal, Walmart, Kering, Publicis, and Virgin Voyages, among others. Seeing it work for both a solo creator and an enterprise brand in the same week still surprises me a little.

The question I get most often is about pricing, specifically how we handle AI given its central role in the product now. The answer is straightforward.

The AI Agent is fully included in Pro. No add-on, no separate usage fee.

Many tools charge extra for AI features. We made a deliberate decision not to do that.

And if you're discovering Inrō for the first time today, we're running 50% off for the first 2 months as part of this launch.
Worth trying now if you've been curious.

If you have questions about pricing, how the plans work, or how Inrō fits into a broader marketing setup, drop them here, and I'll answer directly.

Hey, Pierre here, CTO.

Kshitij covered the what. I'll give you a bit of the how and why, since I know this community tends to care about that.

The hardest part of building the AI Agent wasn't the AI. It was making it behave predictably inside a platform with thousands of active automations already running.

Getting intent detection, mid-flow CRM writes, and agent handovers to work reliably without touching existing flows took most of the last year.

A few things worth knowing under the hood:

1️⃣ AI models tend to be very chatty and quite slow. That does not work for DMs and comments. It took a lot of engineering and optimisation for the AI Agent to reply in a few seconds, including emojis, likes, etc. and still stick to the knowledge provided.

2️⃣ The knowledge base runs on RAG, so the agent retrieves relevant context before responding rather than making things up.

3️⃣ The MCP server exposes 40+ tools to external agents, meaning Claude or ChatGPT can query contacts, trigger scenarios, and manage campaigns directly without touching the UI.

📺 On the product side, the thing I'm most proud of is how much of this runs without configuration. Opt-out detection, spam filtering, folder exclusions, intent triggers.

It all works in the background. You set up your flow and it handles the edge cases itself.

If you have questions about the stack, the API, the MCP integration, or anything technical, I'm here.

Hey, Giulia here from the marketing team. 🙌🏻

I want to flag something for anyone who might want to work with Inrō beyond just using the product.

We run three programs and they don't get talked about enough.

❤️ The Creative Partner Program is for creators and affiliates who want to promote Inrō to their audience and earn from it.

👥 The Agency Program is built for agencies running Instagram across multiple client accounts, there are specific tools and pricing structures that make that setup a lot more manageable.

🤝 The Referral Program is open to all Pro users and lets you earn commission just by recommending Inrō to people you know.

I've been on the marketing side long enough to see this go from a niche automation tool to something enterprise brands and solo creators use in completely different ways for different goals. That breadth is something I didn't fully anticipate when I joined.

Full details on all three programs are at inro.social. If you have questions about partnerships, co-marketing, or anything on the growth and marketing side, happy to talk it through here 😃

This looks intriguing, been searching for something like this to UP my instagram game. Curious to know can we bring our own API key?

AI on SoMe is risky business tho. your account gets banned so fast. And there is no getting your account back after your blocked. Its stricly against SoMe policies etc. Esp on IG. How do you make sure your users dont get banned from using automation?

@kshitij11 Congratulations. And happy product launch.

Curious about the handoff + handback loop. Most just escalate and leave the agent out of it forever. How do you configure when it hands off — is that rule-based or does the AI decide?

Hey Product Hunt 👋 — Marvin here, co-founder of Clawdi.

The honest origin story: I spent three weeks configuring an OpenClaw agent. Skills, cron jobs, 20+ app connections, memory, the whole thing. It was finally working exactly how I wanted. Then a new agent framework dropped that looked genuinely better. I switched.

Everything was gone. Three weeks of work, starting from scratch.

That's the problem we built Clawdi to solve. Every time you switch agent frameworks — and you will switch, because this space is moving insanely fast — you lose everything. Your API keys, your skills, your cron schedules, your agent's memory of how you work. It's not just annoying. It's the reason most people give up on personal agents after a few weeks.

We realized the issue is that your environment — your connections, memory, and config — is trapped inside the framework. So we decoupled them. Clawdi is the environment layer that lives above the framework. Your Gmail, Slack, GitHub, cron jobs, and agent memory live in Clawdi. The agent framework is just the engine you swap in and out.

Switch from OpenClaw to Hermes? Your entire setup carries over. When the next great framework ships next month, you won't start over again.

We also run every workspace inside an Intel TDX hardware-encrypted VM — because a personal AI agent has the keys to your entire digital life, and that deserved real security, not just a promise.

We built this because we needed it. 5,000+ people have tried it since February. Would love to hear what you think — especially if you've felt this pain yourself.

Try it free: https://clawdi.ai

My name is Maggie. I’m the marketing manager at Phala, the TEE infrastructure that Clawdi runs on. So I’ve been close to this project from the start.

My job involves running GTM, social, lifecycle campaigns, content, and a bunch of other things at once. So it's fair to say I am constantly bouncing between tabs, juggling multiple workflows, trying to pulling info from different sources and keep things moving. Also I’m not a developer. I don't have a technical background, so when OpenClaw blew up and because I was actively looking for something to automate my workflows, I tried setting it up locally twice, got lost in the Docker setup both times, and eventually gave up.

Being at Phala meant I knew early on what the team was building with Clawdi. I also knew what it was supposed to solve. But I didn’t actually use it until it was ready enough to just work. I logged in with Google, connected Gmail and Telegram, hit deploy. All under three minutes. My agent said hi to me on Telegram.

First thing I asked Clawdi to do: wiped my Promotions folder. I'd been receiving alerts that my inbox storage is running low but I’d been putting that off for months. 25,500 emails gone in two minutes. Then I set Clawdi up for work. I asked it to pull KPIs from Analytics, the CRM, and our social dashboards every Monday morning and drop a summary in Notion. That used to take me at least 1 hour and a half. With Clawdi all I need to do is to connect it to my work apps with literally just one click and a 1-line prompt “Drop me a clean weekly KPI summary from Analytics, CRM, and socials every Monday morning at 10:00AM EST.”

The part that shaped how we positioned the product: I kept telling the team that the setup was the wall. Not the concept, not the price but the setup. Every non-technical person I showed OpenClaw to bounced at the same point. Clawdi is the answer to that. The 3-minute deployment is NOT a marketing line, it’s the actual fix. With Clawdi integration, the deployment of Hermes agent got even sweeter: Now it doesn’t just do tasks across apps, it learns from every run and gets better over time.

If you work in marketing, ops, or anything that involves a lot of tabs and repetitive tasks, and you’ve already tried OpenClaw and given up, Clawdi is worth another look.

I’ve known the team for a few years. They’ve always been one of the scrappiest teams building in this space. When they told me they built Clawdi on top of OpenClaw, my first question was simple: What can it actually do?

I actually tried OpenClaw briefly when it came out because I heard a lot of AI savvy people had already been using agents to build their "team.” But I gave up after 30 minutes. Well, it was because I didn’t want to spend my time setting things up, not because it wasn’t powerful. Hermes is powerful too, but setup is still the hard part.

Clawdi is where it clicked for me. It took me about 2 minutes to get started, and now I use it for running parts of my Instagram workflow and a bunch of other things. Nothing fancy, just small things that save time every day.

And before anyone asks, yes, your data stays yours. API keys are encrypted and only accessible by you.

These days I’d go to Clawdi just to see what new agents or integrations they’ve shipped, and what else I can offload to my little lobster assistant.

Would love to see how others are using it and what workflows you try on Clawdi.

Hey Product Hunt! 👋 I'm Xiaolu, part of the team at Clawdi, specifically focusing on product research and agent development. It's incredibly exciting to finally share what we've been building with this community.

The honest origin of Clawdi: I kept losing my agent setup. Every time I switched from one framework to another, or picked up a new device, I'd spend the first hour just reconstructing context, re-entering API keys, and re-teaching the agent how I work. It felt like the AI was getting smarter, but the infrastructure around it was still completely fragile.

So we built the layer that was missing. Not another agent, not another framework, just a persistent environment that travels with you. Your memory, skills, secrets, and connected tools live in Clawdi. Every agent you run connects to it. Switch frameworks, add a device, spin up a new agent , your entire setup is already there.

What I'm most proud of: it's MIT-licensed and self-hostable from day one. We didn't want anyone to have to trust a black box with their API keys.

Would love to hear from anyone who's felt this pain, especially if you're running multiple agents in parallel or switching between Claude Code and Codex regularly. That's exactly the use case we've been obsessing over.

Happy to answer any questions below. 🙌

Been waiting for something like this. I've rebuilt my agent setup from scratch 3 times this year switching between frameworks. The "iCloud for AI agents" framing really clicks. Upvoted and sharing this with my team.

Running OpenClaw on Telegram and Claude Code locally — this is literally built for my setup. Installing now.

Congrats on the launch! 🎉

forcing people to share personal API keys or personal agent memory. In other words: team-level vaults, team-level environments, and permissioned access instead of one giant shared account.

We started with personal AI environments first because that pain was immediate, but multi-user / org workflows are one of the next big things we want to support.

This is the missing infrastructure layer. The agents get all the attention but the environment they run in is what makes or breaks long-term usability. Great insight from the Clawdi team.

That's a legitimate concern—abstraction layers only work if they reduce complexity, not add to it. Have you looked into their uptime SLA and whether they offer fallback mechanisms or local-first options for critical agents? Understanding their infrastructure resilience would probably be the first thing I'd validate before committing.

Congrats on launch⭐. How does persistence work for memory and API keys between agents?

I like the vision, but I'm a bit wary of adding another abstraction layer on top of my already-complex agent stack. What happens to my workflow if Clawdi has downtime

Genuinely excited about this. Most AI agent tools compete on flashy capabilities. Clawdi is competing on reliability and persistence — which is actually what matters for daily use. Rare positioning.

Installed in 5 minutes, synced my OpenClaw config to the hub, then spun up Hermes pointing at the same environment. Actually worked. This is the boring-but-essential infrastructure the space has been missing.

Solid concept, but "shared memory across agents" needs a strong access control story. If one compromised agent can write to the shared context, it could affect all others. How are you thinking about isolation?

Is there a way to selectively share part of an environment with a collaborator without giving full access? Thinking about pair-programming-style AI workflows.

Hey, do you have plans for team or org accounts? I'd love to let my team access shared environments without sharing personal API keys.

Good idea！～

Genuine q as a builder, when an agent has persistent memory + cron + app connections all running in your cloud, what does the off ramp look like if I want to pull my setup out 6 months in ? Memory state is usually the part that locks you in hardest, hard to replicate elsewhere once its built up context.

Hi everyone!

XChat is now a standalone messaging app built just for chatting on X.

Elon’s vision for a WeChat-style everything app has been clear for a long time. Taking chat out of the main X app and giving it its own dedicated space is a very intentional step, and it has the potential to become a significant new entry point.

That said, trust will probably be the first real test for many people.

The tension is that you still have to sign up with your phone nr. The private msg without ties to your identity is still wide open. Strangly enough.

Wow! It's super cool Zac! We all knew that hade to be the next step but it's good to see it made it real. Now, let's take the most of it. Congrats on this amazing hunting

Agreed on trust being the test. encryption is the easy part. The hard part is people believing it without thinking about it

snappy + multi-step is the hard combo to nail at the same time — most voice models trade one for the other. what's the latency look like end-to-end on a typical multi-turn workflow?

Voice quality and latency in AI agents is one of those things that's invisible when it works and immediately kills engagement when it doesn't. The "snappy responses" point resonates — for any use case where the conversation has to feel natural (customer support, voice-driven workflows, interactive media), hesitation breaks the illusion.

I've been thinking about this in the context of audio content more broadly. I run a podcast on financial modelling called ModeLoop (https://open.spotify.com/show/0m1oR8AyQv17DVpc5MmirG) and the question of how voice AI changes long-form audio is interesting — not just production quality but whether models like this eventually enable interactive podcast-style experiences where listeners can ask follow-up questions.

For the API use case, what's the typical latency for a first-token response in a complex multi-step workflow scenario?

Whisper changed what people expected from open-source ASR. Three years later, the leaderboard looks very different.

What it is: MiMo-V2.5-ASR is an 8B open-source speech recognition model from Xiaomi MiMo, MIT-licensed and available on HuggingFace, built for bilingual Chinese-English transcription across dialects, noisy audio, code-switched speech, and song lyrics.

The problem: most ASR models are benchmarked on clean studio data and deployed into the real world, where audio is noisy, speakers overlap, and people switch languages mid-sentence. The gap between benchmark accuracy and production accuracy is where voice products quietly fail.

The solution: staged training combining large-scale mid-training, supervised fine-tuning, and a reinforcement learning algorithm specifically targeting the scenarios where conventional models break down. Native punctuation from prosody means transcripts arrive ready to use.

What makes it different: on the Open ASR Leaderboard, MiMo-V2.5-ASR posts 5.73% average WER on English, below Whisper large-v3 at 7.44%. On Wu dialect it scores 19.55% vs FunASR-1.5 at 29.08%. On lyrics, 3.95% on m4singer vs Gemini 2.5 Pro at 4.25%. These are not cherry-picked scenarios — they are the hard ones.

Key features:

• Eight Chinese dialects natively supported, including Wu, Cantonese, Hokkien, Sichuanese

• Chinese-English code-switching with no language tags

• Lyrics transcription under accompaniment and pitch variation

• Multi-speaker and noisy environment robustness

• Native punctuation, no post-processing needed

• MIT license, Python API, Gradio demo, self-hostable

Benefits:

• Production-grade accuracy on the audio conditions that actually exist in the field

• One model replaces multiple regional or domain-specific ASR solutions

• Self-hosting eliminates per-call API costs and keeps data on your infra

• Ready-to-use punctuated output cuts one step from every downstream pipeline

Who it's for: ML engineers and voice product teams building bilingual or Chinese-language transcription pipelines who need accuracy that holds up outside the lab.

Open-source ASR has been catching up to closed models for years. MiMo-V2.5-ASR is a data point that the gap is now very small, and in some scenarios gone.

Dialect and code-switching support is the piece that usually gets skipped in ASR research because it's hard, but it's exactly where real-world audio breaks down. Anyone building a voice product for users in multilingual environments (SEA, MENA, parts of Africa) runs into this immediately.

One application that jumped to mind reading this: location-based audio guides. I built a travel app called StoryRoute (https://storyroute.netlify.app/) that lets people explore cities through interactive, story-driven walks. Accurate multilingual ASR would open up a lot for that use case — imagine a guide that understands a question asked in Mandarin mixed with English street names, or local dialect terms for landmarks.

The code-switching capability in particular seems underexplored for tourism and cultural content. Is the model trained on domain-specific vocabulary or more general conversational speech?

Code switching and lyrics are exactly where ASR demos usually fall apart. Hitting both, plus Chinese dialect coverage, makes this feel grounded in real audio instead of benchmark Code switching and lyrics are exactly where ASR demos usually fall apart. Hitting both, theater. How much latency does that add in live pipelines?

Genspark for Excel is an AI assistant that lives inside Microsoft Excel, handling formulas, charts, data analysis, and even web research, all through plain English prompts.

I'm hunting this because the "AI in your workflow" promise usually means tab-switching. This one actually embeds in the tool millions of people already have open all day.

The problem is familiar: Excel is deeply capable but the learning curve is steep. Most users know maybe 10% of what it can do, and they spend hours on formulas they half-understand, charts they can't get right, and analysis they have to Google their way through.

Genspark sits inside the spreadsheet as a plugin and handles the syntax layer. You describe what you want, it writes and inserts the formula. You select a data range, tell it the story you want to tell, it builds the chart. You ask a question about your data, it answers in plain language.

A few things stand out: the formula explainer (paste any formula, get a step-by-step breakdown), and the web research-to-spreadsheet feature, where it searches the web and fills your cells automatically. That second one is genuinely different from anything I've seen in this category.

Best fit for business analysts, ops managers, finance teams, and frankly anyone who uses Excel daily but isn't a power user.

What's the one Excel task that takes you the longest right now? Are you able to get Genspark to cover?

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

You've nailed why this matters—the embedded experience is the actual innovation here. Most people aren't avoiding Excel, they're stuck in it. The formula explainer sounds particularly useful since understanding what you're looking at is half the battle to actually using it next time.

How does it handle really large sheets? Curious if performance changes once you're past 10k+ rows

Hey Product Hunt 👋

I'm Ivan, maker of PromptPaste.

Like a lot of people here, my best AI prompts were scattered - half in
Notion, half in Notes app. Every time I needed one, I'd either dig for five
minutes or just rewrite it from memory (usually worse than the original).

Every prompt manager I tried was a Chrome extension or a web app, which
meant another account, another tab, and my prompts sitting on someone
else's server.

So I built PromptPaste: a private prompt library that lives in your Mac
menu bar and syncs to iPhone/iPad via iCloud.

What it does:
- ⌘⇧P anywhere on Mac → pick a prompt → it's on your clipboard
- Dynamic prompts with {{variables}} - fill the blanks, copy the result
- Folders, search, favorites
- Share a collection with a link (the only thing that ever touches a server, and only if you choose to)

What "private" actually means:
No account. No sign-up. No telemetry. Your prompts live on
your device and sync through your own iCloud - I never see them. If
PromptPaste shut down tomorrow, your library would still be on your Mac.

Pricing:
7-day free trial, then $59.99/yr, or $149.99 lifetime.

Two things I'd genuinely love feedback on:
1. Where do your prompts live today? (Notion? Notes? Chat history? A Google Doc?)
2. What's the one prompt you wish was always one click away?

Thanks for taking a look 🙏

The local-first, no-account approach is the right call for a prompt library. Prompts are increasingly part of someone's actual intellectual workflow — the way they think, research, and create — and handing that to a cloud service feels like more exposure than most people want.

The folder + variable template system is what makes this genuinely useful rather than just a clipboard manager. For anyone creating content in a specific niche, reusable structured prompts with swappable variables is a huge time saver.

I run a YouTube channel on financial modelling called Mod3Loop (https://www.youtube.com/@Mod3Loop) and a good chunk of my content workflow is prompt-driven — research, script structuring, thumbnail briefs. Having those templates locally organised and instantly accessible from the menu bar would fit that workflow perfectly. Does PromptPaste support syncing across devices via iCloud, or is it intentionally kept device-local only?

Harmony conversations and Codex CLI session logs are not readable by default. That is a problem when your agent just did something unexpected.

What it is: Euphony is an open-source browser tool from OpenAI that renders Harmony JSON/JSONL files and Codex session logs as structured, interactive conversation timelines.

Problem: AI agents running on Harmony-format data produce deeply nested JSON with role tokens, channel markers, and interleaved tool calls. Codex CLI writes a rollout JSONL file to disk after every session. Neither format is designed for human inspection. Debugging agent behaviour means scrolling through hundreds of lines of raw structured data with no visual hierarchy.

Solution: Euphony loads that data from a URL, local file, or clipboard paste and renders it as a readable timeline. Filter by role, recipient, or content type using JMESPath. Inspect message metadata. Edit in-browser. Translate using your own OpenAI API key. All processing stays client-side in the default frontend mode.

What makes it different: It ships as embeddable Web Components, so you can drop a conversation viewer into React, Vue, Svelte, or plain HTML with a single custom element. No need to build a viewer from scratch. The optional FastAPI backend adds Harmony tokenization rendering, which shows exactly how a conversation will be tokenized before it reaches the model.

Key features:

• Load from clipboard, local file, or public HTTPS URL

• JMESPath filtering by role, recipient, and content type

• Metadata inspection panel for annotated datasets

• In-browser editing, focus mode, and grid view

• Embeddable Web Components, Apache 2.0 license

Who it's for: AI engineers working with gpt-oss models, Codex CLI, or Harmony-format datasets for training, evaluation, or agent pipelines.

Agent debugging has needed proper tooling for a while. Euphony is a small but real step toward that.

Hey Product Hunt 👋

The problem we solve:

Every engineering team eventually needs to design, review, and document cloud architecture. Today that means stitching together 5-6 tools: a diagramming tool, spreadsheets for cost, separate security reviews, docs in Confluence, and compliance checklists in yet another system.

What Architecto does:

One platform where you describe what you want to build and get a reviewable architecture.

→ Design: Architect AI generates cloud architectures from natural language. Flow IQ handles flowcharts, UML, sequence diagrams, ERDs, and 4 other diagram types. DB Visualizer covers data architecture.

→ Review: Cost Estimator, Threat Analyzer (STRIDE), Scalability Analyzer, Security Posture, Compliance Checker (SOC 2, GDPR, HIPAA, PCI-DSS), and Drift Detector all run against the same architecture.

→ Document: CoDocs generates docs from code repos. HyperDocs creates visual-rich design documents with AI.

→ Govern: Cloud Discovery, IaC Orchestration, Anti-Pattern Detection, and Architecture Diff keep everything aligned.

Free plan available - 4 diagrams, 200 AI credits, 240+ architecture patterns, and the full learning system (1,000+ quizzes, 20K flashcards).

Would love your feedback. What's the first thing you'd want to try?

The fragmentation problem you described is real and I felt it building DocMetrics alone. Diagramming in one place, cost thinking in a spreadsheet, security considerations in my head, documentation always last and always incomplete. The tax of switching between tools is not just time it is context. Every switch breaks your thinking.

To answer your question directly — the first thing I would try is the Drift Detector. Not because it is the most exciting feature but because it solves the most painful silent problem. Architecture decisions made in week one slowly drift from what is actually running in production and nobody notices until something breaks. Having something that flags that drift continuously is the kind of thing that prevents the 2am incident six months later.

The compliance checker covering SOC 2 and GDPR in the same workflow as design is genuinely smart for startups trying to sell to enterprise. Those compliance requirements usually show up as a surprise after you have already built something and retrofitting them is painful.

One honest question — how does Architecto handle architectures that evolved organically without ever being documented? Most real systems were never designed top down. They grew. Does Cloud Discovery work backwards from what is actually running to create the baseline?

Hey everyone 👋

Smallbits started as a question: what's the minimum number of pixels an icon needs to still be recognisable? Turns out, 8 pixels on each side is the answer – and every point needs to earn its place!

I started this set last summer, kept putting off launching because something was always missing. Eventually I just shipped it. It's a work in progress and I know some categories are thin – but I'd rather hear what you need than sit on it for another year.

A few things worth knowing:

• Completely free, personal and commercial use

• Available as a Figma file and SVGs

• Drawn on iPad in Pixaki, vectorised in Figma

• More icons are coming

If there's an icon you'd love to see added, drop it in the comments. I'm building a list.

Love the constraint - 8x8 forces the kind of clarity that bigger icons let you cheat your way around. The pixel aesthetic also fits perfectly with the retro/terminal-style UIs that are everywhere now. Free for commercial use is super generous too. Bookmarking!

I love this iconset, so good!

The Kafka integration for handling deployment jobs is a smart move—async job processing is exactly what separates stable self-hosted solutions from ones that break under load. The security-first approach with databases bound to localhost is the kind of detail that matters when developers are trusting you with their infrastructure.

Super interesting problem statement. Love to see how the product evolves to solve it.

Yeah feels like there’s a gap here tbh. Right now we’re mostly just manually testing / eyeballing outputs, which isn’t ideal. Let’s try plugging this into our workflow and see how it goes.

I tried it yesterday and it works great! Really like the product. I would probably start using this in my main workflow now.

This is a genuinely smart problem you're solving — the gap between shipping speed and security visibility is real, especially as AI coding tools make it easier to move fast without understanding what's actually in the codebase. The multi-agent scanning approach sounds solid, and explaining vulnerabilities in plain English rather than security jargon will definitely resonate with your target audience.

Hi, I’m Mayurdhvaj, founder of LAEYR.

Most existing tools either try to generate full songs for you or behave like complex DAWs that don’t actually reduce day to day friction. From what I’ve seen, producers don’t want music made for them, they want help with specific parts of the process like coming up with MIDI ideas, generating stems, and iterating faster without breaking their workflow.

That’s exactly where LAEYR fits in. It’s an AI powered assistant designed to sit inside your production process, helping you move faster, explore ideas, and stay in control of your sound.

It’s currently in beta, with more features on the way. I would love for you to try it out and share your feedback, it will directly shape what we build next.

Thank you for everyone who visited azmth! I hope you guys enjoy it, there's a lot more to come <3

Hey everyone! I’m Rhonda.

I’ve spent my life as a Realtor in Saskatoon, so launching a tech startup at 50 wasn't exactly on my radar. But honestly? I was just tired of my phone winning every single night.

I tried the built-in wellbeing stuff on my phone, but it was too easy to cheat. I’d blow my whole “daily limit” by lunch and then just hit “ignore for today” or delete the app because I was frustrated. Traditional blockers felt like a time-out for adults, and I didn't want to be punished—I just didn't want to be stuck.

So I built Paced to handle things differently. Instead of a daily budget, it breaks your time into sessions. When you hit your limit, Paced interrupts the trance and gives you a choice:

• Wait out a 60-minute timer.

• Walk it off (every 100 steps knocks 3 minutes off the clock so you unlock the app faster).

• Emergency Unlock: If you’re desperate, you can get 15 minutes immediately, but it breaks your streak. I’ve found that most people will actually go for a quick walk just to keep that streak alive.

Since I don't have a tech background, I used Gemini as an expert partner to help me navigate all the things I didn't even know I didn't know. It was my sounding board for the strategy and helped me lead a dev team to get this to the Play Store in just seven months. We’re seeing a 30.4% daily usage rate now, which tells me I wasn't the only one looking for a "negotiation" rather than a "jail sentence."

I’ll be here all day to chat about habits, solo founding at 50, or why I think we need a bit more "Wait vs. Walk" logic in our lives.

Paced is on Android now, and the iOS waitlist is officially open!