Most business owners are renting tools they could own. Paying monthly fees to platforms they can't fully control, can't customise and can't build equity in.
I decided to do something about that. I built Relavo - an AI chatbot management platform - primarily because I needed it for my own business. But from the start, I built it with something else in mind too.
I built it multi-tenant. That's a technical way of saying I built the infrastructure to support multiple businesses running their own assistants from a single platform. Not because I needed that for myself - I didn't. But because I could see that what I was solving for myself, other business owners probably needed too. The question was always when, not if.
featured
Build your AI Assistant with Relavo
Join 500+ founders who are vibe-coding their customer support with Relavo.
Try Relavo FreeThat question is what this series is about. Relavo is live, it's working and I'm using it every day in my own business. I already give access to it as part of the premium tier of Vibe Coding Lab - so in one sense, monetisation has already started. But bundling a tool as a community perk and building a standalone SaaS business around it are two very different things. That's the gap I'm trying to close... and if I'm honest, I'm still very much figuring out whether that's the right move at all. This series is that process, in real time, with no airbrushing.
What is Relavo?
Relavo is an AI chatbot management platform. It lets businesses create, train and deploy AI-powered chat assistants without needing any technical knowledge. Think of it as a self-owned alternative to platforms like Chatbase or Tidio, or even custom GPTs - but with a key difference. There are no third-party platform fees eating into your margins. You either bring your own AI model API key or use Relavo's managed models on a pay-as-you-go basis.
It's a multi-tenant SaaS product - meaning multiple businesses run their own assistants from a single shared platform at app.relavo.co.
Here's what it actually does:
You can build and train assistants on uploaded files, scraped URLs, Q&A pairs and free-text content. You can deploy them as an embeddable widget on your website, a standalone chat page or on a custom domain. There's human handoff built in, lead capture, hours of operation, GDPR tools, analytics and Zapier and Make.com integrations. There's also an AI Expert assistant - a premium full-page assistant designed for those who want to sell access to their knowledge via a gated chat interface.
It's a proper platform. Not a prototype.
Why I built it this way
The platforms I was using - and I've tried a few - all had the same problem. You get limited control over the underlying technology, you pay a premium on top of the actual AI costs and you work within whatever constraints the platform has decided to impose. The tool works until it doesn't, and when it doesn't, there's not much you can do about it.
Relavo is built around a different philosophy. Users can bring their own API keys for Claude, OpenAI, Gemini or Grok - which means they're paying for AI at cost, not at a markup. Those who prefer a managed option can use Relavo's models via a wallet top-up system, with transparent per-token pricing. Either way, the control stays with the user.
That's the standard I'd want from any platform I use. So it's the standard I built to.
What it actually costs to run
Here's the honest breakdown at current low volume:
Railway (hosting): ~£5-15/month
Supabase (database): free tier for now, Pro plan ~$25/month as it scales
Resend (transactional email): free up to 3,000 emails/month
Stripe (payment processing): 1.4% + 20p per UK card transaction
Cloudflare (DNS and CDN): free for most usage
Total fixed infrastructure: roughly £50-80/month.
The managed AI model costs - for users who don't bring their own API keys - are designed to be self-funding. Users top up a wallet in £10, £25 or £50 increments and per-token charges are drawn from that balance. At the right margins those costs pay for themselves.
Pricing is correct at the time of writing and subject to change.
The pricing model
I built the subscription tiers with a clear logic: start accessible, grow with the business. Having been on the other side of platform pricing that felt designed to push you up a tier as quickly as possible, I was deliberate about making sure each plan gives you something real.
Solo: £15/month - 1 assistant
Starter: £29/month - 5 assistants
Growth: £49/month - 15 assistants
All plans include a 14-day free trial. Cancelled accounts drop back to a limited free state rather than losing access immediately - which feels like the right thing to do. Nobody should lose everything they've built because they need a month off.
Whether these specific tiers are the right structure for a standalone product is something I'll be working through in this series. The pricing philosophy won't change. The numbers might.
Where things stand right now
Relavo is fully built and functional. But it's pre-launch in the sense that the Stripe live price IDs still need connecting and a couple of domain configurations need finishing. The platform works. The billing isn't live yet.
That's what Part Two of this series will cover - the pre-launch checklist, getting the first paying customers through the door and what that process actually looks like.
This series isn't going to be a highlight reel. It's going to be the real thing - the decisions, the numbers, the things that work and the things that don't. I don't have all the answers yet. That's the point.
Relavo was built using the vibe coding workflow I teach inside Vibe Coding Lab. If you want to build your own tools, come and join us.
