Who owns the key, history, and model bill
TypingMind also appears in our broader OpenRouter alternatives guide for buyers comparing BYOK clients with bundled model access. We also compare Capriole AI with Poe, Magai, and Venice.
Capriole includes the model access
Capriole takes responsibility for the access layer. Signed-out visitors and Free users can first try supported flagship models in free browser chat. Premium then provides unlimited web chat on every supported model. Public API and coding-agent traffic draw from 5 million included monthly charged API tokens. That replaces several administrative steps: opening provider accounts, storing several keys, funding several balances, and checking several billing pages. The browser workspace, API allowance, and maintained agent paths sit under one Capriole account. TypingMind’s personal product works differently. You pay for the interface, add a provider API key, and continue paying the selected provider for model usage. Optional cloud storage can add another cost after the included allowance. For example, chatting 5M tokens a month with GPT-6.2 Astra at three input tokens per output token costs USD 80 at OpenAI’s standard API rates, billed by OpenAI on top of the TypingMind license. In Capriole AI, that chat is part of the USD 9 Premium plan. Capriole users do not need separate provider keys. The one AI subscription guide explains the included access and the boundary between unlimited browser chat and metered API work.Capriole gives storage choices without changing products
Capriole offers three chat-history choices. Synced saves account history. Local keeps history in the current browser. No Storage leaves no saved chat history. Every mode still sends the prompt through Capriole and the chosen model provider. The chat storage guide explains how to use synced history, browser-local history, and temporary chats. Free image generation with GPT Image 2.6, Nano Banana 2.1, and Seedream 5.5 is also included for signed-in accounts that use Server Storage; see the image generation guide.One workspace can continue into external tools
Capriole lets a user continue one active conversation with another supported model while keeping the existing thread context in the workspace. The same account then extends into OpenAI and Anthropic protocol endpoints and 10 documented coding-agent paths. Capriole AI provides the workspace, supported model access, public API allowance, and external-agent configuration together.Capriole’s API is part of the personal workspace
Capriole’s public endpoints are intended for applications and agents that speak OpenAI Chat Completions, OpenAI Responses, or Anthropic Messages. One Capriole key and API balance sit behind those supported paths. TypingMind Teams does have a Custom API for its team workspace. Personal custom endpoints work in the other direction: they let TypingMind call a compatible model service supplied by the user rather than including an external model gateway.BYOK moves the cost and setup back to the user
TypingMind offers BYOK, self-hosting, MCP, plugins, and deep client customization. Those options still leave provider accounts, keys, balances, and model bills with the user. Capriole includes supported model access and carries the workflow from browser chat into API and coding-agent use under one account.Use Capriole without assembling the stack
Choose Capriole when you want supported frontier models, unlimited Premium browser chat, an included API allowance, coding-agent paths, and storage controls without assembling the access layer yourself. Try the free browser chat to use supported flagship models without first creating provider accounts, buying an interface license, or supplying separate API keys. For more, upgrade to USD 9 Premium for unlimited chat and an API key. Capriole AI is an independent workspace, not affiliated with TypingMind.Official sources
- TypingMind documentation
- TypingMind general FAQ
- TypingMind personal licenses
- TypingMind personal and team comparison
- TypingMind Team plans
- TypingMind Team API overview
- TypingMind feature list
- TypingMind privacy policy
- TypingMind cloud sync and backup
- TypingMind multi-model responses
- TypingMind MCP documentation

