Any agent CLI
Works with Claude Code, Copilot CLI, Cursor, Aider, or any command that accepts a prompt. Configure per-machine or per-task.
Parallel agent output
Every node runs its own local agent and writes hostname-scoped results, so you can compare outputs side by side.
Zero infrastructure
No backend, scheduler, database, or inbound port. OneDrive handles coordination. One command to start.
How it works
Submit a task once. Each running machine sees the immutable task file, executes locally, and writes back a separate result file.
Machine A Machine B Machine C
agentfleet run agentfleet run agentfleet submit
│ │ │
└──────────┬───────┴──────────┬────────┘
▼ ▼
OneDrive / AgentFleet shared workspace
tasks/ → immutable task JSON
output/ → per-machine results
Try it on real agent work
Multi-agent code review
Ask every machine to inspect the same repository for security, reliability, or architecture risks.
Distributed refactor rehearsal
Let separate machines attempt the same migration on separate clones, then compare patch quality.
Prompt benchmark lab
Run one prompt across different local agents, models, branches, or hardware profiles.