What is Calca?
Calca is an AI-first agentic design tool. That sentence packs a lot, so let me break it down.
Most design tools today treat AI as a copilot — something you invoke, prompt, and then take its output to continue your work. Calca inverts that model. In Calca, the AI agent is the primary actor. You set the goal, the AI acts, and you review the result. It’s a subtle but profound shift in how humans and AI collaborate in creative work.
Works with Any AI Model
One of the founding principles of Calca is model agnosticism. Whether you prefer OpenAI, Anthropic, a local model via Ollama, or any other AI provider, Calca can work with it. There’s no vendor lock-in, no forced abstraction layer. You bring your own API keys, your own models, and Calca adapts.
This means:
- You can experiment with different models for different tasks without switching tools
- Privacy-conscious users can run local models entirely on their own hardware
- Teams can choose cost-effective models for routine tasks and premium models for complex decisions
Desktop-First
Calca runs as a native desktop application. No browser tab to lose, no SaaS subscription to maintain, no internet connection required for core functionality. Your projects live on your machine.
Being desktop-first isn’t a limitation — it’s a philosophy. Creative work benefits from sustained focus, rich local integrations, and the ability to work offline. Calca respects that.
Free to Use
Calca is free. Not a trial, not a freemium tier with artificial limits — free. I believe agentic design tools should be accessible to everyone, not just teams with large budgets.
The Big Picture
Calca is more than a design tool. It’s a bet on a future where AI agents handle the mechanical work of design — iteration, variation, layout — while humans focus on intent, taste, and judgment.
In the coming posts, I’ll dive deeper into how Calca is built, the architectural decisions that make it possible, and the challenges of bootstrapping an agentic tool without massive infrastructure scaling.
Repository: https://github.com/espetro/calca