Quora's Poe launches Previews, letting users create web apps like data visualizations and games using more than one LLM, like GPT-4o and Gemini 1.5 Pro
Poe, Quora's subscription-based, cross-platform aggregator for AI-powered chatbots like Anthropic's Claude and OpenAI's GPT-4o …
Context & Ripple Effects
Poe had been building an ecosystem around third-party AI experiences: it introduced per-message pricing and analytics for bot creators after Quora said it would use new funding to expand Poe and pay creators. Previews extends that arc from conversational bots toward user-made software.
The feature matters because Poe sits above multiple model providers rather than tying creation to one assistant. That positioning becomes more useful as users choose models for different tasks rather than treating model access as a single-product decision.
First-order effects
- Poe users can create lightweight web apps, including visualizations and games, with access to more than one LLM such as GPT-4o and Gemini 1.5 Pro from within Poe.
- Poe gains a new creation surface alongside its bot marketplace, potentially giving its existing creator tools a broader class of products to support.
Second-order effects
- Multi-model app creation shifts more of the user experience—and the choice of which model to invoke—to Poe’s interface, while model providers compete to be selected inside those apps.
- Creators who previously monetized individual bots can test app-like experiences as another way to differentiate, building on Poe’s creator pricing and analytics rollout.
Third-order effects
- If multi-model creation becomes a standard feature, AI aggregators may evolve from chatbot directories into workflow and application layers that mediate access to underlying models.
- The durable competitive question shifts toward creator distribution, tooling and cross-model orchestration; model quality still matters, but it is no longer the only product surface users encounter.
The trend: AI chatbot aggregators are moving from model comparison toward multi-model creation environments that package several LLMs into user-facing workflows and apps.