Everything is around the task
A task is a container for intent, context, and action. Not a to-do item — a living, structured primitive that connects every person, device, and AI agent through a single shared layer.
ImagineTask is built on two deeply integrated engines. The Task Engine handles any task. The UI Engine renders it — in real time, for any actor.
The core runtime of ImagineTask. It accepts any task you can describe — structures it into a typed container of intent, context, and action — then routes, executes, and evolves it. No task is out of scope.
There is no fixed interface. The UI Engine reads a task's structure and state and renders the right interface — in real time, for whoever is looking at it. Every actor sees exactly what they need, when they need it.
What needs to happen and why. Expressed by a person, triggered by a device, or initiated by an AI agent — the intent is the task's purpose.
Everything surrounding the intent. Who is involved, what environment it operates in, what constraints exist, and what has already happened.
What gets done. The executable steps, delegations, triggers, and events that carry the intent through to completion.
A task is not a record. It is not a message. It is a living container that carries full context across every handoff — between people, devices, systems, and AI agents — without losing anything along the way.
When context travels with the task, every actor — human or machine — can pick it up and act on it immediately. No reconstruction. No gaps. No lost thread.
Hand off work with full context intact. No more reconstructing intent from scattered messages — the task carries everything the next person needs.
Direct machines, sensors, and services through structured tasks. The device receives not just a command, but the full context of why and how.
Machine-to-machine coordination through a shared task layer. No brittle point-to-point integrations — every node speaks the same language.
Tasks are the grounding layer for AI agents. Agents read state, act on tools, and update context — always working within a structured, auditable container.
When a system needs a human decision, it surfaces a task — with full context, history, and options — so the person can act without having to investigate.
Tasks compose. Parent tasks spawn children. Subtasks report back. Complex workflows are expressed as task graphs — not code, not config.
ImagineTask is not an AI agent. We provide the structured task layer that agents need to do real work — a typed, auditable environment where agents can read intent, act with context, and report outcomes without ambiguity.
Retrieve the full task container — intent, context, fields, connections, and current action state.
Update context, add fields, or progress state. The UI Engine re-renders automatically for all connected actors.
Execute a named action within the task's sandbox. Chain, delegate, or escalate — all within a structured boundary.
Stream real-time task state changes. Agents stay in sync with the task as it evolves — no polling, no guessing.
ImagineTask is in early access. If you're building systems where people, devices, and AI agents need to work together — this is the infrastructure layer you've been missing.