Give your infrastructure an operations agent. It finds problems, plans fixes, and carries out approved work.
Hyphen Agent investigates problems, right-sizes workloads, and cleans up across clouds and clusters. It works continuously within your policies, with evidence and approval for sensitive changes.
Every change follows the same loop. Propose. Approve. Execute. Verify.
- 01
Connect, read-only
Agent starts with the same permissions as the member using it, and infrastructure access begins read-only. It inventories and observes before any change is possible.
- 02
Plan updates backed by evidence
Findings arrive as plans: what it found, why it matters, the confidence, and the exact change it wants to make.
- 03
You approve
Sensitive scale and cleanup actions wait for a decision. Approve once, set a policy, or say always ignore. Repository changes go through pull requests.
- 04
It executes, then verifies
Approved work runs with preflight checks, post-change verification, rollback options, and a complete audit history.
The recurring work, off your plate.
Agent investigates, recommends, and carries out work within your policies or with your explicit approval.
Findings, not log spelunking. Agent analyzes cloud provider logs and New Relic APM data, surfaces findings by severity, and explains likely causes and affected resources.
From finding to reviewable fix. Agent creates GitHub issues from findings, comments on matches, and proposes fixes through pull requests you review and merge.
Flags don’t outlive their launch. Agent detects Toggle flags with no recent usage or unchanged values over a 90-day window and opens cleanup pull requests.
Signals investigated before they page you. Agent investigates deployment and runtime signals, models scaling options, and recommends or executes changes within your availability targets.
Catch stale resources before they become waste. Agent detects orphaned containers, expired preview environments, and scaled-to-zero workloads, then queues cleanup for your review.
Every action, on the record. Ask Agent to list, count, and summarize activity across deployments, environments, flags, and tasks. Every action it takes lands in the audit log.
It’s your infrastructure. Agent works inside your rules.
Set organization-wide policies for your Agent tasks, then let teams customize them by project, environment, cluster or workload.
One platform. Every place you run.
Your cloud accounts
Connect AWS, Azure, or Google Cloud. Agent observes, diagnoses, and optimizes what’s already running, including applications Hyphen didn’t deploy.
Bring your own cloudKubernetes, anywhere
Public cloud, private cloud, on-premises, or bare metal. Clusters connect outbound through Horizon, with no public management endpoint.
Kubernetes managementPut your operations agent to work.
Connect your infrastructure and let Agent investigate, optimize, and carry out approved work within your policies.