Slava Akulov

We store all our company's data in GitHub. So, Claude Code and Codex plus MCP are part of our daily routine. The problem of amnesia is real. I have a technical question: Why are you using blockchain to store memory? What's the advantage? @sebsim

We're also building accounting automation for SMBs. Our main channel is WhatsApp and Messenger, where we want an AI accountant to chat through Messenger. So, it's a cool idea. I've been waiting for a solution like this for a long time. We'll definitely check it out and test it.

I work on a Mac and always use Caffeine for this, but it would be awesome if it was native when I need to work. When the agents finish their tasks, Caffeine doesn't stop, so this is actually a good idea if it can be done. Thanks for the product, I think it looks really cool.

We manually maintain CLAUDE.md across our whole monorepo, it's a real struggle. When building multi-agent architecture, especially when embedding context or passing feedback, everything just gets lost and you can't restore the session. So this looks like a pretty interesting idea.

Hey Product Hunt! I'm Slava, CEO of Jupid. I co-founded Anna Money (100K+ users, CNBC top UK fintech). At Anna Money I was COO — ran everything from product to finances. We had 15,000 active clients running on autopilot with just one accountant. I've been automating accounting for 5 years — first for Anna Money, then for my own companies. When Claude and GPT-4 arrived, I made four decisions: I...

@alice_viola_setti I was just yesterday looking for a solution to visualize agent work. And it seems like... I searched in Claude, searched in ChatGPT, and each of them offered their own solution that didn't really satisfy me. So listen, great idea. I think this definitely needs to be tried, especially for overnight runs.

We're a team of 6 building on GitHub as a shared OS. Agents, skills, finances, communications, content, investor decks, all in one place. Everyone reuses the same agents across tasks. The "learns how you work" part got my attention, but I'm curious whether MuleRun is built for one person or if there's a team layer. When the whole workflow lives in one place and you're sharing agents across...

I keep running into this and can't figure out a good way to handle it. Claude shows "thinking" but it's unclear whether it's actually working or stuck in a loop, maybe hitting some architecture issue underneath. Is there any detection for that in Masko? Like, can it tell the difference between "model is processing" and "model is retrying silently and going nowhere"?

The 'learns how you work' angle is what caught my attention — I've spent years building automation tools and the hardest part is always making them context-aware without manual setup. How does it handle domain-specific workflows that combine different tool stacks? Curious whether it can track patterns across Claude Code sessions and terminal commands, which is where most of my work happens.

I've been using Claude Code with MCP for about a year now and keep iterating on my setup. Curious what's in Garry's stack — is it mostly about the MCP server configuration or also specific prompt engineering patterns? The biggest unlock for me was getting file system + browser MCP tools working together. Does GStack cover the agentic workflow side or is it primarily IDE-focused?

MCP integration in 2 lines is a strong hook — the real friction in generative UI isn't rendering components, it's getting the LLM to emit the right structure reliably. We build LLM workflows for structured financial data and the jump from "returns JSON" to "returns interactive UI" is where most teams get stuck. Curious: how does C1 handle edge cases where the model's UI intent is ambiguous — do...

I spend a serious amount of time hunting down and wiring up Claude Code skills for our accounting automation stack — discoverability is the real friction. Most skill repos are scattered or undocumented. Centralizing this is the right call. Curious: do you curate for quality before listing, and is there a way to pin to a specific skill version so a production workflow doesn't break when a skill...

As someone who runs a remote team of 5 across different time zones, unclear Slack messages are my daily pain. We waste real time re-reading and guessing intent. Love that you're attacking this — it's one of those "obvious in hindsight" problems. Does it work in real-time as you type, or does it analyze messages after sending?

This is exciting — I use MCP daily with Claude Code for automating financial workflows, and seeing it expand into design tooling feels like the right next step. The idea of AI agents working directly inside Figma rather than just describing changes is a big unlock. Curious — how granular is the control? Can agents handle component-level edits, or is it more layout-level for now?