Why Radioso

Using the internet is about to mean talking, not clicking.

For thirty years, using the internet has meant clicking through websites. That’s changing: more and more, it means talking — to agents that book the flight, find the recipe, fill the grocery order, and do a hundred things nobody has imagined yet. Every organization will need a way to take part in that conversation — to answer, to help, to do — wherever its customers are already talking.

And here’s what we believe: as a handful of large agents come to mediate more of that conversation, every organization should be able to represent itself — on its own terms, in its own words, not through whatever summary a big model happens to hold about it. That belief, not a feature list, is why Radioso is built the way it is.

• Open source and self-hosted. Run it yourself, read exactly how it answers, shape it to fit. Every conversation stays in your Postgres.
• Grounded in what’s true about you. Answers come from documents you publish and cite their sources, so anyone can check the work. No evidence, no invented answer — and if you can break that, we want the bug report.
• Steered by rules you author. The agent works out how to help; your directives and routines decide what it may do while it does. When it gets something wrong, you see which rule steered it and fix that — no vendor ticket, no retraining.
• Your model keys, no inference markup. Bring OpenAI, Anthropic, Gemini, or any OpenAI-compatible endpoint. You pay your provider, not a middleman.

Talking is the cheap half

The job is finishing the thing the person came for. A Radioso agent takes real action instead of describing it: routines carry a return request or an onboarding check across turns, skills call your tools and webhooks, and when the moment calls for a person, the agent hands over the whole conversation with everything it did marked inline. It knows when to act — and when to ask. That’s the difference between talking about it and getting it done.

Nobody clicks a product you don’t sell

People tell an assistant things they would never put in a search box. In our experience, 0.5–2% of visitors talk to the assistant when it’s offered — at 50k monthly visitors, that’s about 500 conversations a month where people tell you, in their own words, why they came. On one site, visitors kept asking the assistant for course dates instead of browsing course pages: they already knew the course, they just needed the details. At a web store we work with, several customers asked about a perfume category the store didn’t sell. Guess what, they do now.

Radioso ships that loop built in. Audience Pulse turns the last 30 days of visitor questions into a census — named topics with exact counts, recurring gaps where answers had nothing to stand on, and content recommendations you can open as a draft document. The full argument is on our blog: Know thy customer .

One deployment, every surface

Your website needs a chat widget. Your sales team wants a shareable link. Your support tooling calls the API, and your engineers want the same agent in Cursor and Claude Desktop. That usually means several integrations, several contracts, and several copies of your knowledge drifting apart.

With Radioso you build the agent once — its data, its rules, its persona — and every surface talks to that one agent, with the same answers and the same debug loop. Update a document or a routine, and every surface updates with it.

Whether we’re the right fit

We’d rather lose a deal than win the wrong one, so here is where the edges are.

Knowledge you curate, plus systems you call. The agent answers from material you publish into Radioso — docs, policies, product pages, crawled sites — and grounding is only as current as what you have published. Live state that changes minute to minute, like ticket status or inventory, belongs on the other side of a skill: you connect the system over MCP and the agent queries it at the moment of the question. Deciding which knowledge is curated and which is fetched live is a real design call, and it is yours to make.

Agents you put in front of people. Every part of this product is shaped around a conversation someone is actually having: a persona, standing rules, a flow that survives across turns, a handoff to a teammate. If the thing you need is a workbench for assembling arbitrary AI pipelines that no human talks to, you want a tool built for that job.

You own and operate the stack. Radioso is open source, runs on your own infrastructure, and keeps every conversation in your Postgres. That also means you run it: upgrades, capacity, model keys, and the on-call pager are yours. Some teams want exactly that. Some teams would rather pay someone else to hold it, and that is a fair trade to want.

If those trades read as the ones you would make, keep going.

Get started

Every organization should shape its own conversations. Here’s where yours starts: