Per-agent, per-run
Attribute every dollar to the exact agent, workflow, and execution that generated it, making every cost fully traceable.
The Economic Control Plane for AI Agent Fleets
Your agents spend money on every call, tool, and token. Xerg shows you where it goes, what it wasted, and what to change.
Three monetary findings. Seven neutral signals. No inferred waste.
Copy/paste to your agent and go:
Audit data stays on your machine unless you choose to push it.
See all setup paths →
How it works
One path.
Free to prove, simple to scale
Install locally, run an audit that returns real numbers, then create an account when you’re ready to act on them.
From install
to governance
Economic control plane
Spend analyzed: $148,392
See spend as a map, not a bill
A bill tells you what you already lost. Xerg maps spend to the agent, route, and action that caused it, so waste is a place you can point to, not a number you find too late.
Routes and tools
Identify which models, tools, API calls, and execution paths contribute the most to your AI spend and optimization opportunities.
Policy drafts
Turn confirmed waste into draft spending policies and recommended fixes, each validated with a before-and-after comparison of your actual runs.
FAQ Frequently
asked questions
Still have questions?
Whether you’re evaluating Xerg or planning
a larger deployment,
we’re here to help.
Is Xerg safe to install?
Yes. Xerg runs locally and only analyzes the information required to generate your audit. It doesn't take control of your agent or access unrelated data.
Do I need an account to get started?
No. Install the skill and run your first audit locally, free, with nothing to sign. An account is only needed when you want to save audits, track spend over time, and share the map with your team.
Which AI agents are supported?
Xerg reads OpenClaw, Hermes, QM, Claude Code, and Cursor usage data. QM uses a platform-independent sanitized snapshot adapter with strict direct or certified Fly-contained collection. Any other framework can send events through xerg ingest.
What does the audit show?
A spend map of every call, tool, and token, attributed to the agent, workflow, and run that generated it, with waste classified in dollars and concrete fixes recommended.
How long does the audit take?
Minutes. The audit reads the logs your agents already produce, so there is no instrumentation to add and no code to change.
Can Xerg be used by enterprise teams?
Yes. The Enterprise plan adds SSO and audit logs, custom retention, and priority support with an SLA for organizations running AI at scale.
What is included in the paid plans?
Team, Growth, and Scale all include full audit history, unlimited pushes and members, Ask Xerg, hosted MCP, Slack notifications, and policy drafts. The band depends on monthly audited agent spend, not seats.
Ready when you are Point your agent at
Copy/paste to your agent and go:
Audit data stays on your machine unless you choose to push it.
See all setup paths →