Changelog | ChatAcademia

Claude Sonnet 5 for Pro Chats

ChatAcademia Pro now uses Claude Sonnet 5 for stronger academic reasoning, tool use, and long-context research workflows.

Route Pro chats through Claude Sonnet 5 while keeping the existing Fast and Max tiers available
Keep Pro chats resilient with Gemini and OpenAI fallbacks when an upstream model is temporarily unavailable
Update credit estimates and provider settings to match current Sonnet pricing and adaptive thinking

More Reliable Restricted-Browser Sessions

ChatAcademia now handles browsers that block local storage more gracefully, keeping chat sessions stable when privacy settings are strict.

Fall back safely when a browser allows storage reads but blocks saved preference writes
Keep analytics setup from interrupting chat pages in privacy-restricted Firefox sessions
Refresh supporting app, API, website, and editor dependencies

Stripe-Backed Trials and Plans

New workspaces now choose a Stripe-backed trial or subscription plan after onboarding, with clearer pricing and billing controls.

Replace the legacy free plan path with a seven-day Basic trial for new and eligible legacy workspaces
Show monthly and annual plan choices with clearer pricing, credits, and subscription actions
Keep subscription, cancellation, trial, lifetime, and daily credit reset state aligned through Stripe webhooks and billing sync
Keep direct chat links and annual billing prompts stable after a trial is canceled or a workspace has no active plan

More Reliable Chat History Recovery

Older research chats now recover more clearly when a saved message history can no longer be replayed by the current AI model runtime.

Show the existing recovery flow when an incompatible chat history fails during streaming
Reduce confusing low-level model message errors after reopening older research chats
Keep agent streams from reporting expected incompatible-history failures as production crashes

Newer Academic Chat Models

ChatAcademia's research agents now use newer model routing and a more direct AI SDK tool runtime for stronger academic responses.

Route Fast, Pro, and Max chat modes to refreshed OpenAI, Gemini, and Claude model choices
Improve tool execution across research agents by using the current AI SDK runtime context directly
Keep credit estimates aligned with current no-cache and long-context model pricing

Changelog ChatAcademia