Calcis estimates LLM API costs before you send a request. Paste a prompt, select a model, and see the projected cost including predicted output tokens. Unlike observability tools that show costs after API calls are made, Calcis gives you the estimate upfront so you can make decisions before spending anything. Key features: - Exact token counts using each provider's native tokeniser - Output token prediction trained on millions of real prompt/response pairs - Multi-turn conversation cost simulation - Agentic workflow cost estimation - GitHub Action for automated PR cost comments - Public REST API for CI/CD integration - Model recommendation engine - Supports 24+ models across OpenAI, Anthropic, and Google - Free to use, no account required for basic estimates