Cloud agents — questions and answers

Running coding and research agents on cloud machines that keep working after you close your laptop.

Each question links to the page that answers it. The answers are not repeated here — they belong to the page that earns them.

70 questions across 15 pages · browse the Cloud agents pages →

• Is Polaris an official cloud version of Claude Code?

  Cloud agents that keep working after you close the laptop

• Can a Polaris worker check out my repository and run my tests?

  Cloud agents that keep working after you close the laptop

• What happens if the runtime machine dies mid-job?

  Cloud agents that keep working after you close the laptop

• Do I need my own Anthropic API key?

  Cloud agents that keep working after you close the laptop

• How is any of this priced?

  Cloud agents that keep working after you close the laptop

• Can I point Polaris at my own machine instead of a hosted one?

  Running an AI agent in the cloud instead of on your laptop

• Which model runs the session?

  Running an AI agent in the cloud instead of on your laptop

• Does the agent see my whole workspace?

  Running an AI agent in the cloud instead of on your laptop

• How fast does work start after I assign it?

  Running an AI agent in the cloud instead of on your laptop

• What does a session cost?

  Running an AI agent in the cloud instead of on your laptop

• Does my browser need to stay open for the agent to keep working?

  What keeps running after you close the lid

• What if the job was mid-search when I closed the lid?

  What keeps running after you close the lid

• Is partial work lost when a job fails?

  What keeps running after you close the lid

• Can I see what happened while I was away?

  What keeps running after you close the lid

• Do AI workers count as seats on the bill?

  Turning a single-player agent into a team member

• Can two people assign work to the same worker at once?

  Turning a single-player agent into a team member

• Can a worker create tasks for humans?

  Turning a single-player agent into a team member

• How does a teammate know an agent is working right now?

  Turning a single-player agent into a team member

• Does everyone in the org see every agent's work?

  Turning a single-player agent into a team member

• Can my teammate change how the worker behaves?

  Sharing the agent, not the transcript

• How does a colleague ask for a revision?

  Sharing the agent, not the transcript

• Can I keep some agent work private?

  Sharing the agent, not the transcript

• Does sharing a worker mean sharing my API key?

  Sharing the agent, not the transcript

• Can I schedule a recurring agent task?

  Queueing work at night and reading it in the morning

• How many jobs run at the same time?

  Queueing work at night and reading it in the morning

• What if a job fails at three in the morning?

  Queueing work at night and reading it in the morning

• Will I be billed for a night that produced nothing?

  Queueing work at night and reading it in the morning

• Can I run a coding agent on my phone with this?

  Running your AI workers from a phone

• Do I need to install anything?

  Running your AI workers from a phone

• Does voice input change what the worker receives?

  Running your AI workers from a phone

• Will I be notified when a delivery arrives?

  Running your AI workers from a phone

• Is a cloud agent slower than a local one?

  Cloud agents versus local agents

• Can a cloud agent read my codebase?

  Cloud agents versus local agents

• Which is safer?

  Cloud agents versus local agents

• Do I have to pick one?

  Cloud agents versus local agents

• What limits a single cloud session?

  Cloud agents versus local agents

• What if I assign a task with no checklist?

  Assigning a task to an AI worker

• Can I change the task after assigning it?

  Assigning a task to an AI worker

• How do I stop a session I regret?

  Assigning a task to an AI worker

• Can I assign the same task to a person and an agent?

  Assigning a task to an AI worker

• Does the worker see the whole bucket?

  Assigning a task to an AI worker

• Can an agent delete its own history?

  The audit trail behind every agent session

• How long is the session stream kept?

  The audit trail behind every agent session

• Does the audit trail record what the model was thinking?

  The audit trail behind every agent session

• Is any of this exportable?

  The audit trail behind every agent session

• Who can see an agent's work in a private bucket?

  The audit trail behind every agent session

• How is a skill different from the worker's instructions?

  Skill files, and why a worker's capability should be readable

• Can I edit a skill file?

  Skill files, and why a worker's capability should be readable

• Do skills consume context on every job?

  Skill files, and why a worker's capability should be readable

• Can a worker gain a skill during a session?

  Skill files, and why a worker's capability should be readable

• Can I connect a tool that is not in the catalog?

  How an AI worker gets tool access

• Is the credential per worker or per organisation?

  How an AI worker gets tool access

• What stops a worker using a connection it should not?

  How an AI worker gets tool access

• Does the agent need my Anthropic key too?

  How an AI worker gets tool access

• How do I revoke access?

  How an AI worker gets tool access

• What counts as a round?

  How long an agent session can run, and what happens when it ends

• Does a retry start from scratch?

  How long an agent session can run, and what happens when it ends

• Can several jobs run at the same time?

  How long an agent session can run, and what happens when it ends

• What if the model keeps calling tools and never delivers?

  How long an agent session can run, and what happens when it ends

• Am I billed for a session that hit its limit?

  How long an agent session can run, and what happens when it ends

• Does the agent decide whether to produce a file, or do I?

  When the deliverable is a file

• Are the PDFs real PDFs?

  When the deliverable is a file

• Can a worker revise a file it already delivered?

  When the deliverable is a file

• Where do the files actually live?

  When the deliverable is a file

• Can it produce a spreadsheet or a slide deck?

  When the deliverable is a file

• Can an AI worker ever close a task?

  Agents deliver, humans close

• What happens to comments the worker could not address?

  Agents deliver, humans close

• Can I get the previous version of a file back?

  Agents deliver, humans close

• What does a rating do?

  Agents deliver, humans close

• Does asking for changes cost more?

  Agents deliver, humans close