The smart routing feature automatically selects the most appropriate model for each task based on the nature of the query. Users can also manually switch between models mid-conversation without losing any context, which means conversation history, uploaded files, and the thread of discussion all carry forward when jumping from one AI provider to another. This context preservation matters when comparing outputs or when one model struggles with a particular request and you want to immediately try another without re-explaining everything.
Real-time cost tracking displays the price of each message before it gets sent. Every interaction shows its exact cost, giving users granular visibility into their AI spending at the message level rather than discovering totals at month's end. This transparency extends to the entire workspace, where users can organize conversations into folders by project or topic, search across all chats instantly, and export sessions as PDF or markdown files.
This service generates files directly from conversations. Users can create markdown documents, JSON data, code files, and other structured content within chats, then download and share these artifacts with team members. The artifacts feature isn't available on the Economy plan, though.
Three pricing tiers offer different token allowances and model access. The Economy plan costs $9.99 monthly and includes 1 million tokens with access restricted to OpenAI and DeepSeek models only. Smart routing works at this tier. The Professional plan jumps to $29.99 monthly with 3 million tokens and expands model access to OpenAI, Anthropic, Google, and xAI, while excluding Mistral and Perplexity. It adds the artifacts feature. The Expert plan costs $99.99 monthly and provides unlimited tokens under a fair use policy, meaning excessive automated usage might require moving to a custom enterprise plan. This tier enables all six providers including Mistral and Perplexity, plus a Knowledge Ledger feature that only exists at this level.
A free trial lets users test the service before committing. The company operates as Xerotech Ltd and maintains registration with the UK Information Commissioner's Office under number ZC065188. They emphasize that private data and personally identifiable information never reach the AI models themselves.
The workspace includes team sharing capabilities, letting multiple users collaborate on conversations and access shared artifacts. The search function works across the entire conversation history, making it practical to find specific exchanges or information from past chats without manually scrolling through folders.
Professionals juggling multiple AI tools for different strengths represent the core audience. Someone who subscribes to one service for coding help, another for writing tasks, and a third for research can consolidate into one interface while maintaining access to specialized models for each use case. The cost tracking appeals to users who need to justify or monitor AI expenses, particularly those billing clients or tracking departmental budgets.
The token limits on lower tiers constrain heavy users. One million tokens disappears quickly with extensive daily use, and the Economy plan's restriction to just two providers limits the multi-model comparison that makes this service useful in the first place.
