The Evolution of Calendar Systems

iCalendar solved a specific problem: synchronizing event records between systems. Start time, end time, title, location, recurrence rules, attendees — a portable text representation that Google Calendar, Apple Calendar, Outlook, and others could exchange. For scheduling meetings and sharing availability, it worked well enough to become the default.

The format assumes events are relatively static records. Edit the file, re-sync, propagate changes. Discovery, rich media, conversational context, and model-grounded suggestions were not design requirements — they were not imaginable requirements when the spec was written.

Three gaps matter for planning products built in the language-model era:

• Multimodal context. Notes, photos, links, and model conversations do not have a native home in .ics. They live in separate apps, detached from the event they describe.
• Retrieval and discovery. There is no schema for indexing event context, searching across timelines, or grounding model responses in structured history. Calendars store appointments; they do not store memory.
• Dynamic interaction. Natural-language scheduling — propose, preview, revise, commit — requires mutable state and human-in-the-loop approval flows. A static text record is the wrong primitive for agent collaboration.

Calendar infrastructure is entrenched. Billions of devices read and write .ics and CalDAV. Replacing it outright is neither realistic nor necessary. The evolution we care about is additive: a layer above legacy interchange that preserves interoperability while introducing time-indexed context, structured events, and model-grounded workflows.

Óra integrates with calendar systems where users already keep commitments. Events can export and sync through existing channels. The planning environment itself — timeline, notes, images, prompts, contextual memory — lives in a richer structure that .ics was never meant to carry.