Poe

Poe, built by the team behind Quora, took a different bet from most AI assistant products: instead of building a single model interface, it became a marketplace where users could access Claude, GPT-4, Gemini, Mistral, Llama, and dozens of custom bots in one unified interface. Launched in 2022, it attracted millions of users who wanted to compare model outputs without managing separate subscriptions to every AI provider. For AI power users and developers prototyping bot behaviors, Poe became a surprisingly important aggregation layer.

The technical challenge Poe faces is unique among consumer AI products: it must maintain stable, low-latency connections to multiple third-party AI APIs simultaneously while providing a consistent user experience regardless of which underlying model is responding. When any upstream provider — Anthropic, OpenAI, Google, or Meta — experiences latency spikes or API issues, Poe's users feel it on that specific bot. Poe's own infrastructure handles authentication, message routing, subscription billing, and bot hosting for creator-built custom assistants, each representing an additional failure surface beyond the big model providers.

When Poe has server issues, the symptoms vary by component: the web app or mobile app failing to load the conversation interface, specific model bots returning "bot is unavailable" errors while others work fine, messages sending but responses never arriving and the typing indicator spinning indefinitely, subscription purchases failing to process or premium access not unlocking after payment, custom bot creation tools returning errors when saving, and push notifications for new bot responses not arriving on mobile.

Outage.gg tracks Poe's platform status alongside its major upstream model dependencies. When a specific bot stops responding, the status page can help you determine whether the problem is with Poe's infrastructure or with the underlying model provider it routes to.