Fleet 1.13:Teams are now shipping 5x more PRs with autonomous pipelines.See what's new →
FleetFleet
Comparison

Fleet vs. Google Jules: Self-Hosted Fleet Orchestration vs. Google's Async Cloud Agent

Jules is Google's cloud-hosted async coding agent tied to GitHub and Google's model stack. Fleet is a self-hosted orchestration layer that coordinates multiple role-based agents with model flexibility, governance, and no dependency on any single cloud provider.

Google Jules is an async AI coding agent that works on GitHub issues in the background, producing pull requests using Google's Gemini models. It runs entirely on Google's infrastructure, requires no local setup, and has been generally available since August 2025 with free and paid tiers.

Fleet runs on your infrastructure and coordinates a full roster of Claude Code agents — developer, reviewer, release manager, PM — each with its own role and run-time budget. It does not depend on Google, Microsoft, or any cloud provider. The fabric event bus handles handoffs between agents so the pipeline runs autonomously end-to-end.

Choose Fleet if

Teams that want self-hosted, multi-role agent orchestration with governance — independent of any cloud AI provider's roadmap.

Choose Google Jules if

Developers who want to assign GitHub issues to an async AI agent backed by Google's Gemini models with no local infrastructure to manage.

Fleet vs. Google Jules: side by side

FeatureFleetGoogle Jules
DeploymentSelf-hosted binary, any machine or serverGoogle cloud service, no local install
Agent runner / modelRuns Claude Code as the agent runnerGoogle Gemini (proprietary)
Agent rolesDev, reviewer, release manager, PM, QA, and moreSingle coding agent
Agent handoffsFabric event bus with reactive chainNot supported — one agent per task
Data residencySelf-hosted; your source code goes only to your model backend (Bedrock/Vertex in your own cloud) + GitHubProcessed on Google infrastructure
Audit trailFull decision log and fabric event historyGitHub PR comments
Budget controlsPer-agent run-time budgets; 6-dim evaluation + auto-quarantine risk modelNot exposed to operators
AvailabilityGenerally available at fleetctl.aiGenerally available since Aug 2025; free / AI Pro / AI Ultra tiers

Where Fleet is the better fit

  • Self-hosted — your source code stays on your infrastructure (runs the model via Bedrock/Vertex in your own cloud), which matters for security-sensitive codebases
  • Full multi-role pipeline: issue to dev to reviewer to release manager without manual handoffs
  • Governance stack: run-time budgets, 6-dimension evaluation, and a separate auto-quarantine risk model
  • Not tied to any single cloud AI provider's roadmap — runs Claude Code on your own infrastructure

Where Google Jules is the better fit

  • Zero setup — no binary to install, no YAML to write; connect GitHub and assign issues
  • Backed by Google's model investment, which may improve task quality on complex coding problems over time
  • Native GitHub integration for issue assignment and PR creation without any webhook configuration
  • Good for teams that prefer fully managed infrastructure and have no hard requirements on data residency

Pricing

Jules is generally available with a free tier and paid access via Google AI Pro ($19.99/month) and Google AI Ultra (higher tiers). Fleet offers a free tier (1 agent slot), Team at $49/slot/month, and Enterprise pricing. For a 5-role fleet, Team costs $245/month plus model API costs.

Do they compete, or coexist?

The tools do not directly interoperate, but Fleet's watcher can react to the same GitHub label changes that Jules would respond to — so teams could route some tasks to Jules and others through Fleet's full pipeline, depending on complexity.

Frequently asked questions

Is Jules available to use today?

Yes. Jules has been generally available since August 2025, with a free tier and paid access through Google AI Pro and AI Ultra subscriptions. Fleet is also generally available at fleetctl.ai.

Why would I choose Fleet over Jules if Jules requires no setup?

If your requirements include self-hosted data residency, multi-role agent pipelines, or governance controls, Jules's simpler setup comes at the cost of those capabilities. Fleet gives you control over the entire stack.

More comparisons

Fleet vs. Copilot Workspace →Fleet vs. Devin →Fleet vs. SWE-agent →

Run your first agent fleet

One binary. Five minutes. See every agent, coordinate every handoff, and keep a full audit trail of what your fleet did.

See how it worksInstall Fleet