CrbonFree

CrbonFree helps you understand how much AI you use across your AI chat sessions. The extension tracks token usage across supported AI chat platforms and coding tools, giving you a consistent view of your AI activity across services. CrbonFree currently supports: ChatGPT — estimated token usage for regular and Work conversations Claude — estimated token usage for supported Claude conversations Gemini — estimated token usage for supported Gemini conversations Claude Cowork — exact token usage when usage data is available from the service Claude Code — exact token usage when reported directly by Claude Code For browser-based AI conversations where providers do not expose exact token counts, CrbonFree calculates a local estimate based on the conversation content. These values are clearly presented as estimates and should not be treated as provider billing totals. When exact usage information is available from a supported tool, CrbonFree records the reported token counts instead of estimating them. CrbonFree lets you: View AI token usage while you chat Track usage across multiple AI platforms Review usage from individual conversations See your AI usage history in one place Organize usage by account, organization, or project when signed in Distinguish estimated usage from exact provider-reported usage CrbonFree is designed to make AI usage easier to understand without changing how you use ChatGPT, Claude, Gemini, or supported coding tools. Privacy-first AI usage tracking Guest Mode When you use CrbonFree without signing in, token tracking runs 100% locally in your browser. No account is required, and your usage data is not sent to CrbonFree servers. Token counting and estimation happen entirely on your device. Signed-in Mode When you choose to sign in, CrbonFree sends your AI usage information, such as token usage, to the CrbonFree platform so you can maintain a long-term usage history across sessions and devices. Your private chat messages are not sent to or stored by CrbonFree. Conversation content may be processed locally in your browser when necessary to estimate token usage, but the message content itself is not uploaded to our platform.