RUNTIME LABS
Coordinate Agents ・ Schedule Actions ・ Manage Memory
Prompt & Plan
Use natural language to store memories, schedule events, and adapt your plans just in time — including images, location, time, and context — all within a unified calendar and timeline.
Image & Voice
Share photos, notes, and reflections with events. With Óra's image-centric timeline, your calendar becomes more than a static archive — it's a visual expression and a window into past memories and future plans.
Runtime Labs
Runtime Labs is an independent lab led by Derek Rosenzweig.
Óra is our first release. Feel free to reach out with questions, feedback, or thoughts.
Time
Schedule Events
Manage and coordinate tasks using a calendar-based runtime. Schedule plans aligned with your goals.
Temporal Reasoning
Interpret time in context. Ground plans in schedules, recurring events, and evolving user intent.
Memory
Interoperable Memory
Edit, transfer, and structure memory traces across agents. Maintain continuity and adaptation through user context.
Feedback & Reflection
Capture feedback to improve model behavior. Enable agent learning through reflective trace editing and summarization.
Autonomy
Accountability
Audit and monitor agent behavior over time. Enable inspection of planned and executed actions for transparency.
Agency & Control
Preserve user control by exposing structured plans, traceable logic, and feedback loops that support ongoing alignment.
Agency emerges from our capacity to predict how actions shape outcomes across time.
With transparency in memory and planning, we gain autonomy to select trajectories aligned with our goals.
— Runtime Labs
Chronologue
A time-based reasoning and calendar interface that grounds model behavior in time. It exposes a programmable context layer from calendar events and supports conversation-driven schedule updates for dynamic planning, reflection, and agent coordination.
MemPort
A memory logging and reflection vault designed for long-term memory traces and journaling. It integrates with Chronologue to enable structured feedback, summarization, and agent adaptation over time.
Tempo Tokens
A duration-aware tokenization layer that encodes time, preference, and scheduling constraints. It underlies Chronologue's reasoning stack and enables plan evaluation grounded in user intent and temporal alignment.
Agent Runtime
A structured execution environment for coordinating agent behavior. It manages plan queues, logs execution traces, and exposes FastAPI endpoints for model orchestration and external runtime control.