All activity
Mykola Kondratiukleft a comment
the clickbait problem is real. i stopped opening most news apps because the headline says "major breakthrough" and then it's just a study with 12 participants and a huge disclaimer at the bottom. appreciate the straight-forward approach here. curious how you rank importance - is it algorithmic based on source/engagement, editorial judgment, or some mix? also wondering if you filter by topic or...
One Minute NewsLearning what happened around the world in one minute.
Mykola Kondratiukleft a comment
honestly surprised macOS still doesn't have this built in. i keep spotify pinned in one desktop and slack notifications blast at full volume during focused work sessions. curious about the Focus Audio auto-duck feature - does it detect calls from multiple apps (zoom, meet, discord) or do you manually configure which apps should trigger it? also wondering if it persists per-app levels across...
VolumeHubPer-app volume control for macOS with no driver installs
Mykola Kondratiukleft a comment
i have like 30 tabs open at any given time - one for docs, one for the API playground, one for Stack Overflow from two days ago that i swear i'll read eventually. the "one tab you never close" angle is smart. what made you go with browser tab vs desktop app? i'm guessing it's for the zero-friction start, but curious if there were trade-offs around performance or persistence when the tab...
StayAwake.DevThe one tab developers never close
Mykola Kondratiukleft a comment
been running Claude Code and Cursor side by side for months, biggest pain is when they both want to touch the same file. the worktree approach is smart - does each agent get its own isolated git state or do they somehow sync changes back? also curious how you're handling permissions, like can agents accidentally blow away each others work or is there some kind of conflict detection built in?
AxelTodoist for AI coding agents
Mykola Kondratiukleft a comment
honestly this has been my biggest headache building AI tools. I've got a few products out there and the pricing model question keeps me up at night. went with tiered SaaS for simplicity but the problem is heavy users absolutely crush your margins while light users subsidize them. the hybrid approach makes the most sense to me - base tier for access + usage for compute-heavy features. but the UX...
Is usage-based pricing becoming the norm for AI tools?
Jake FriedbergJoin the discussion
Mykola Kondratiukleft a comment
agent identity is going to be a huge problem as more autonomous agents start hitting APIs. the reputation score concept is interesting - how do you handle cold start for new agents? do they need to build up trust from zero or is there some way to inherit reputation from their operator?
NervePay Give AI agents identity, secrets vault & analytics
Mykola Kondratiukleft a comment
the voice shortcut customization is the killer feature here. I've been vibe coding with Claude Code for months and honestly the biggest time sink is re-typing similar prompting patterns over and over. being able to say P1 and have it automatically add planning mode context? that's the dream. curious how the codebase linking works - does it just read file names or does it actually understand the...
CyphrKeyWisprFlow for vibe coders w/ voice shortcuts
SomeYum just leveled up! Now you can plan your entire week and generate smart grocery lists with one tap.
🆕 What's New:
🗓️ Weekly Meal Planner - Pick recipes for each day, organize by meal type, track completion
🛒 Smart Grocery Lists - Auto-generated from your meal plan, organized by store section, quantities calculated
Still the same swipe-to-save recipe discovery you love - now with the tools to actually execute your meal plans.
SomeYumPlan your week's meals, generate grocery lists instantly
Mykola Kondratiukleft a comment
Hey PH! 👋 Built SomeYum to solve my own dinner decision fatigue - but realized saving recipes is useless if you never actually cook them. So I added meal planning + grocery lists. Now I plan Sunday, shop once, and actually cook what I saved. The grocery list auto-categorizes by store section and calculates quantities from recipes. No more wandering the aisles. Would love your feedback - what...
SomeYumPlan your week's meals, generate grocery lists instantly
Mykola Kondratiukleft a comment
100% yes. My most-used product started because my wife and I couldn't decide what to cook for dinner. That's it. A recipe swiper app. Boring problem, but it's something millions of people deal with daily. Meanwhile the "sexy" AI products in my portfolio get way more attention on Twitter but way less actual daily usage. Boring problems = recurring pain = people actually come back. Flashy...
Are the best startups built on boring problems?
Boris GostroverhovJoin the discussion
Mykola Kondratiukleft a comment
Claude (Sonnet 3.5/3.6) for speed, Opus for complex architecture decisions. I've shipped 4 products as a weekend vibe coder over the last 3 months using almost exclusively Claude Code. The thing nobody mentions: the model matters less than your prompting discipline. Clear specs, small scoped tasks, reviewing every output. I write maybe 30% of the code myself but review 100%. The security gap is...
What's the best AI model for coding?
fmerianJoin the discussion
Mykola Kondratiukleft a comment
The workspace separation is a smart design choice. Every all-in-one writing tool eventually becomes a mess because blog writing and novel writing have completely different mental models. Having dedicated tools for each but under one roof avoids the "I need 5 apps" problem. The blog workspace with SEO + readability scores is what would get me - currently juggling between my writing app and...
Scribeist V2Write without switching tools
Mykola Kondratiukleft a comment
This solves a real problem. I've been running Claude Code for months and the "just trust it" approach to file system access is... uncomfortable. Especially when you're working across multiple projects and one agent could theoretically touch files from another. The sandbox-exec approach is smart - native macOS sandboxing without the Docker/VM overhead. How does it handle when Claude Code needs...
MultituiSandbox claude code, codex, or any TUI on macOS
Mykola Kondratiukleft a comment
The "near zero false positive rate" from Matt's review is the real differentiator. Every AI code review tool I've tried drowns you in noise - 50 comments per PR where maybe 2 matter. If you've actually solved signal-to-noise, that's huge. Curious about the security angle - does it catch things like hardcoded secrets or auth bypasses that slip through when AI-generated PRs come in faster than...
Unblocked Code ReviewAI code review that knows when to chime in
Mykola Kondratiukleft a comment
The "fragmentation tax" is real. I've been vibe coding 4 apps as weekend projects and the pattern is always the same: build in 2 hours, spend 6 hours on deployment/infra/auth glue. Having that collapsed into one workspace could be a game changer. Question for the team: any plans to include security scanning in the pipeline? When AI generates the code and you deploy in one click, there's a real...
CreateOSBuild and deploy apps from any AI coding tool, in one place
Mykola Kondratiukleft a comment
Different tools for different vibes honestly. Claude Code when I want to explore a quick idea or refactor something small. Cursor when I need full project context - it handles multi-file changes way better. Biggest pain with both: they'll happily add dependencies without checking if they're maintained or have known CVEs.
Cursor or Claude Code?
Gabe PerezJoin the discussion
Mykola Kondratiukleft a comment
Been bouncing between Cursor and Claude Code lately. Cursor for longer sessions where I need project context, Claude Code for quick scripts. My hot take: the "best" tool depends on whether you're prototyping or building something you'll maintain. For throwaway projects, speed wins. For production code, I'd add a security scan step before deploy - AI loves to add dependencies it doesn't vet.
What is the best Vibe Coding tool so far? Bonus points if we've never heard of it!
Gabe PerezJoin the discussion
Mykola Kondratiukleft a comment
This is exactly what vibe coding needs - the sandbox approach is smart. Been seeing too many people run coding agents with full system access, not realizing what can be exfiltrated. The network sandboxing idea for preventing secrets theft would be huge. Most security issues I've seen in AI-generated code aren't runtime vulns - they're leaky APIs, hardcoded creds, or packages that phone home.
MultituiSandbox claude code, codex, or any TUI on macOS
Mykola Kondratiukleft a comment
The "fragmentation tax" framing is spot on. I've been vibe coding side projects and the biggest friction isn't writing the code — it's the deployment dance afterward. Claude writes it, then I'm juggling Vercel, Supabase, Stripe dashboards... Curious about the AI-native chat — does it have context from previous deployments? Like if I say "add auth like we did on the last project" does it know...
CreateOSBuild and deploy apps from any AI coding tool, in one place
Mykola Kondratiukleft a comment
Hey Product Hunt! 👋 I'm Nick, and I built VibeCheck to solve a problem I kept seeing: developers using ChatGPT and Claude to build apps, then launching without realizing they have security holes. Why I built this: Last year, I watched several indie hackers go viral... for the wrong reasons: - One had their entire database dumped because of SQL injection - Another had hardcoded API keys that got...
VibeCheckCheck if your AI-generated code is safe to launch