Pricing

Pay for pager classes solved.
Not dashboards rendered.

The agent runs on your cluster. You pay for the managed control plane — runbook orchestration, audit log, escalation fan-out — not per seat, not per host, not per alert. No credit card to start.

Start free — helm install byteport → Talk to founder →
Free / Open
$0
self-hosted · no credit card · forever

Run the full agent on your own cluster. All open-source runbooks included. Community Slack for help. No call-home required for the core detect → diagnose → act → verify loop.

helm install byteport →
  • Open-source agent on your cluster (MIT)
  • Up to 3 active pager classes
  • All shipped signal adapters (Prometheus, CloudWatch, GCP, Datadog, K8s, Sentry, PagerDuty, Grafana…)
  • RBAC scope enforcement + dry-run mode
  • Auto-generated postmortems (local)
  • Community Slack support
  • No managed control plane
  • No audit log export
  • No SSO / SAML
Most popular during beta
Team
$199
per cluster / month · 14-day free trial · no seat fees

All pager classes, managed control plane, audit log export, Slack + PagerDuty escalation fan-out. One price per cluster — no per-seat math.

  • Everything in Free, plus:
  • All shipped pager classes (unlimited active)
  • Managed control plane (high-availability)
  • Audit log export (S3, GCS, Splunk, syslog)
  • Slack + PagerDuty escalation fan-out
  • Email escalation with postmortem attachment
  • Custom runbook authoring (YAML)
  • Spend cap configuration
  • 99.9% SLA on control plane
  • Email support — 1 business day
Enterprise
Custom
on-prem or dedicated tenant · SSO + SOC 2

On-prem control plane, custom IAM scopes, SSO/SAML, dedicated Slack channel, named ops engineer, SOC 2 Type II on the roadmap. For teams that need procurement-grade assurances.

  • Everything in Team, plus:
  • SSO / SAML (Okta, Azure AD, Google Workspace)
  • On-prem control plane — no external egress
  • Custom IAM scopes (AWS, GCP, Azure)
  • Volume runbook pricing
  • 99.95% SLA on control plane
  • SOC 2 Type II report on request (roadmap)
  • Named ops engineer for runbook authoring
  • Custom data retention & compliance controls
  • 24/7 support + dedicated Slack channel
What's included

Free vs Team vs Enterprise

The agent binary is the same on all tiers. What scales is the managed control plane, compliance surface, and support depth.

Feature Free Team Enterprise
Active pager classes 3 Unlimited Unlimited
RBAC scope enforcement ✓ ✓ ✓ + custom IAM
Dry-run / safe mode ✓ ✓ ✓
Audit log export ✕ ✓ S3/GCS/Splunk ✓ + custom
Escalation channels GitHub Issues Slack, PagerDuty, Email + dedicated Slack ch.
Managed control plane ✕ ✓ HA, 99.9% SLA ✓ On-prem option
SSO / SAML ✕ ✕ ✓ Okta, Azure AD
SLA Best effort 99.9% 99.95%
Support response time Community Slack 1 business day 24/7 + named engineer
On-prem option ✓ agent only ~ agent only ✓ full stack
SOC 2 / compliance ✕ ✕ ~ roadmap Q3 2026

Private beta pricing — rates are locked for any team that installs before GA. See how BytePort stacks up against PagerDuty AIOps →

FAQ

Common questions

A pager class is a category of incident the agent knows how to diagnose and remediate end-to-end — for example, disk_pressure, pod_crashloop, or db_connections_saturated. Each class has a runbook: a signal trigger, a decision tree, safety guardrails, and a verify step. BytePort currently ships 10 flagship pager classes — see exactly which signals each class handles and what it does without waking you. Free plans activate up to 3 simultaneously; Team and Enterprise activate all of them.
Yes, on Enterprise. The byteport-agent binary always runs on your cluster — no signal data leaves your infrastructure on any tier. The control plane (runbook orchestration, audit log, escalation fan-out) is managed by us on Free and Team. Enterprise adds a fully on-prem control plane with no external egress requirement, including air-gapped Kubernetes clusters. Email us to start the conversation.
The agent requests only the RBAC verbs it needs to execute enabled runbooks — nothing broader. Every permission is declarative, documented, and auditable. Dry-run mode lets you verify what the agent would do before granting write access. See the full RBAC grant table and IAM scopes on the Security page.
Free tier is free, full stop. Team tier is $199/cluster/month during the beta. Any team that installs before GA has that rate locked — it will not increase when we exit beta. We're not offering hidden discounts; the price you see is the price you pay, locked.
Email us and we'll switch you to Enterprise with prorated billing for the remainder of the month. No service interruption, no re-install required. Team plans don't artificially throttle on runbook volume — the limit is the control plane SLA and support depth, not execution count.
Free tier is the startup plan — it's genuinely free with no expiration. If you need Team features (managed control plane, SSO-adjacent things, audit log export) but you're pre-Series A, reach out. We'll work something out. We'd rather have you running and happy than turn you away on price.
No. Escalations — Slack alerts, PagerDuty pages, GitHub Issues, postmortem emails — are free on all paid tiers. You pay for the managed control plane, not for the times the agent decides a situation is outside safe scope and calls for help. An agent that's never wrong about when to escalate is more valuable than one that charges you for knowing its limits.
BytePort can run alongside PagerDuty (consuming it as a signal source via the PagerDuty adapter) or replace its AIOps layer for teams primarily buying that for alert grouping. If you use PagerDuty for on-call scheduling, keep it — wire BytePort as the remediation layer on top. If you're paying for PagerDuty AIOps and want something that actually fixes the incident instead of just grouping the alerts, Team is the switch worth making. Full comparison →
Open by default
Security & RBAC 60-second install Public changelog Open runbook source GitHub (MIT)

Runbook source is MIT-licensed and on GitHub. RBAC permissions are declarative and auditable. Every action the agent takes is logged. Nothing is hidden behind a SaaS wall.

Start free — helm install byteport

Wire up one signal adapter in under 5 minutes. The agent starts monitoring immediately. No credit card. No call-home. Just a running agent on your cluster.

60-second install guide → Talk to founder →