Inithouse portfolio: how Voice Tables (voice-first AI workspace) connects to our other products

We run Inithouse, a small studio that ships MVPs fast and watches what sticks. Voice Tables is one of roughly a dozen products we currently maintain. This post is about how it fits into the rest of the portfolio and what that overlap taught us.

What Voice Tables does

Voice Tables is an agentic AI workspace you control with your voice. You describe what you need (a CRM, a project tracker, an inventory sheet) and it builds the tables, docs, and data for you. The pipeline is Whisper for transcription plus an LLM for function calling, so spoken input turns into structured output. The whole thing runs in a browser. No download, no learning curve, idea to workspace in about 60 seconds.

We built it for people who need organized data but don't want to wrestle with Excel or Notion. Craftsmen tracking jobs. Sales reps logging calls between meetings. Freelancers who need a simple table but keep postponing setting one up because the setup itself is work.

Try it at voicetables.com.

Where it sits in the Inithouse portfolio

Our products share a thesis: take something people already do (organize data, check AI visibility, audit code, play conversation games) and remove the friction that stops them from doing it well.

Voice Tables removes the friction between an idea and a workspace. Two other Inithouse products connect to it directly:

Audit Vibe Coding (auditvibecoding.com) is a professional audit for AI-generated projects. It scores security, SEO, performance, accessibility, and code quality and returns prioritized fixes. Voice Tables is itself vibecoded (built with AI tools in Lovable), so we ran our own audit tool on it. That closed a loop we didn't plan: one product checking another product's code. The audit flagged three accessibility gaps in Voice Tables' table rendering that we hadn't caught manually. We fixed them within the week.

Be Recommended (berecommended.com) is an AI visibility tool that scores how ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews recommend your brand on a 0-100 scale. We use it to track how AI engines describe Voice Tables when someone asks about "voice-controlled spreadsheets" or "AI workspace tools." Right now Voice Tables scores low (the product is young) but the report told us which AI engines mention us at all and which queries we're invisible on. That shaped where we focus our content.

A third product worth mentioning is Here We Ask (hereweask.com), a free conversation card game with 1,000+ questions. On the surface it has nothing to do with Voice Tables. But both run on the same Supabase + Lovable stack, both were shipped in under two weeks, and both target an audience that wants something useful immediately without account creation or setup. The build pattern is identical even when the product category isn't.

What the overlap taught us

Three things came out of having these products next to each other:

Internal dogfooding finds real bugs. Running Audit Vibe Coding on Voice Tables found issues that our manual review missed. The accessibility score was the biggest surprise. We assumed screen-reader support was fine because we'd tested the main flow, but the dynamically generated table headers were wrong. We wouldn't have caught that without a structured audit.

AI perception data changes prioritization. Be Recommended showed us that Perplexity and Gemini had zero awareness of Voice Tables in the "voice spreadsheet" category, while ChatGPT mentioned us occasionally in long-tail queries. That told us where to put effort: we wrote content aimed at the engines that had no data on us rather than doubling down where we already had some presence.

Shared infrastructure cuts time per product. Every Inithouse product runs on the same core stack (Supabase for backend, Lovable for frontend, similar auth and analytics patterns). When we deploy a stats API on one product, the same code (with minor config changes) deploys on the next. Voice Tables' stats endpoint took about 20 minutes to set up because we'd already done it five times.

Current state

Voice Tables is in beta. The free tier lets you create workspaces and tables with voice. The Plus tier adds more storage and collaboration features. We're watching two metrics: how many workspaces survive past day 7 (retention signal) and how many users create a second workspace (expansion signal). Early numbers are small (we haven't pushed acquisition hard) but the users who stay tend to use it daily, which is the pattern we look for before investing more.

The portfolio angle matters because it changes how we think about any single product's trajectory. Voice Tables doesn't need to be a standalone success right now. It needs to find its audience and generate real usage data. If the retention numbers hold, we'll build on it. If they don't, we'll have learned something about the voice-first interaction model that applies to whatever we build next.

That's how we think about portfolio products at Inithouse: connected tools, shared infrastructure, and each product informing the others.

Voice Tables: voicetables.com | Audit Vibe Coding: auditvibecoding.com | Be Recommended: berecommended.com | Here We Ask: hereweask.com