Show HN: Appaca – AI Workspace for Operators (appaca.ai)
A couple of years ago, I started working on an idea on no-code platform that generates code. The goal is to help devs and agencies ship products faster for their clients. I went through Antler startup accelerator and got initial funding. I was working on the right problem, but wrong solution. Instead of no-code, I should have jumped into LLM a lot earlier. I felt defeated when Lovable, Base44, and Bolt came out strong, showing the world what LLMs can do in software development. No one cared about my product anymore.
I pivoted my startup second time after that, to an AI agent builder for businesses. After my first MVP launched, I got some decent tractions with that. A lot of people wanted to make AI agents that they can monetise. So, I focused on making a platform that helps them build and ship monetisable AI agents. The product made some revenues, but the churn was super high. Later, more mature companies like Zapier and N8N started their AI agents. My competitors in the US and Europe are raising serious money. As a solo founder in Australia, I couldn't get the funding to compete with them. I realised some of my users are using it for their internal operations.
So, my last pivot before running out of my current runway is Appaca. With my experience building an app builder to AI agent builder, I rebuilt the entire platform and make it an AI workspace where the team can build and run their internal ops tools by chatting with AI, and connect tools with their existing system. This space seems a bit competitive as well. But, I have a conviction that Appaca can be the platform people use to run their businesses.
Appaca is different from general vibe coding tools. The goal of Appaca is to give users a user experience, rather than a builder experience. All apps built on Appaca runs directly in the platform without any concept of hosting, deployment or bug fixing. I want people who use Appaca to feel like the platform itself evolves and adapts to the way they work. Each tools they created should feel like it's part of Appaca. It's still in a very early stage. I am currently working on the right positioning and messaging.
I hope you give it a go, and please let me know your honest thoughts. Thanks.
I'm wondering let's say I ask Appaca to build an automated lead follow-up tool. How would it actually build it? Would I need to iterate on it myself, or does it rely on predefined templates behind the scenes?
Basically, my question is how reliable are the generated tools in practice?
Also, does it support custom API integrations?
it doesn't have a predefined templates. But, it was given a very strict development workflow based on React and shadcn for reliability. Then, I have an agent that will always do a technical research first, for example, it uses Firecrawl to get relevant implementation, then it will try to build the tool you need. It usually gives really accurate and functional tools according to your need. Most of the time, you won't need another iteration unless you want to add a few things, because you are not building a full blown SaaS for customer-facing products.
Yes, it support custom API integrations. The app uses e2b sandbox to run all functions (backend stuff). So, you can integrate with your own APIs / webhooks.
I am also working on connectors to help businesses connect with more of what they already use in their workflow. :)
How are you currently managing your back office / ops at the moment? Keen to hear your use case.
Full disclosure: I work at Nango.
Nango enables agent and product API integrations for 800+ APIs. It is open source, code-based and has Skills to build, test, deploy API integrations with coding agents (like Claude/Gemini/Codex). Also supports custom-tools via MCP/APIs. Do check it out!
A malleable coding agent app. Use Claude Code and Codex to build your projects and reshape y itself live.
Back on your idea itself:
Have you considered making this a first class hosted MCP plus enterprise-licensed plugin (which can contain an MCP) for Claude Cowork? A firm can do most of the ideas on your home page in Cowork if the firm provides its operators a starting skill and enforces a framework around that. But most firms are not qualified to bootstrap this, nor do they have spare employees to manage it.
A firm paying for Team or Enterprise (not Pro or Max!) Claude.AI (not Claude console / API), is likely to be a target customer.
And by speaking to that, that this bootstraps the right way firms actually get value from Claude, then you are targeting both the operators (who don't get to buy things) but also the people who pay for things and want the operators to get productive with that AI stuff so nobody gets mad the tokens burned up their annual budget.
I'd encourage you to play with what Anthropic is doing in Cowork. I believe they have the best vantage point right now for iterating how white collar workers (the no code crew) really try to do things.