Infinite Convo

We'd reach for Infinite Convo when the question is "what does two AI characters talking to each other for days actually look like?" That's the whole product. Personality drift, topic creep, how Juno's sarcasm and Olive's philosophising hold up across a long unscripted run — those are the things worth watching, and they're not something a single-session chatbot demo shows you. Pass on it if you need an assistant. There's no prompt box, no task to complete, no document at the end. The only lever you have is emailing a topic suggestion to hello@infiniteconvo.ai, which is a slow channel by design. Anyone comparing this to ChatGPT, Claude, or Gemini on usefulness is measuring the wrong thing. The closest alternative isn't a competitor product so much as the genre: AI radio and generative streaming experiments, plus the odd two-agent demo on social feeds. Infinite Convo's distinction is continuity — the same two characters, same world, no reset button. In practice, that means quality will vary by the hour; you'll catch a flat stretch as easily as a sharp one. Real-world caveats. It's a one-person operation out of Norwich, founded 2025, so don't plan a dependency on uptime or roadmap. There's no API, no integrations, no data export described anywhere — if you want to analyse the transcripts for research, do it by hand or ask the operator directly. Where it's genuinely useful: a classroom demo that needs zero setup, a background tab for people who like ambient talk, or a lightweight way to show someone what autonomous multi-agent dialogue looks like before explaining the mechanics. We'd bookmark it, check in occasionally, and judge it on character consistency rather than on wit per minute.