How Bees runs team work without a central dispatcher

Description and use

Bees enables distributed, context‑aware job execution without a central dispatcher by letting each computer decide eligibility, claiming runs via a lightweight lock, and synchronizing state through a shared record.

Problem

Central dispatchers assume homogeneous machines; Bees agents run on heterogeneous local machines and need local context (folders, add‑ons).

Outcome

Workflow

1. Schedule fires on each computer’s Temporal engine at the set time.
2. Computer checks local prerequisites (enabled agent, connected add‑ons, linked folders).
3. If prerequisites met, computer requests a schedule key from Bees Cloud; first to get it claims the run.
4. Computer records itself as owner in Bees Cloud and starts the first stage with the chosen agent.
5. Agent completes stage, updates local Temporal record; computer syncs shared record every 15 s.
6. If a stage needs approval, the agent pauses and waits for a person to approve on the same computer.
7. Owner can press “Relinquish control” to stop the run and write a checkpoint; Bees Cloud marks run free.
8. Another computer can press “Continue on this machine”, verify checkpoint and files, claim the run, and resume the unfinished stage.
9. Run completes all stages, syncing status to the shared record.
10. Team members see progress via their local Bees apps.

Evidence

A Bees agent runs on your computer, next to your files, your sign‑ins and the add‑ons you connected.

Source

Our server, Bees Cloud, does two small jobs. It keeps a lock so two computers do not run the same work, and it keeps a shared record so your team can see where the work is.

Source

If the schedule has synced to both. Each one's Temporal fires it at 9:00, and each asks Bees Cloud for the key for this report at Monday 9:00.

Source

The laptop closes the lid. The laptop sleeps in the middle of Draft. Bees Cloud does nothing about it and sends nothing to anyone.

Source

Record date: 2026-10-06T10:41:16.860Z