Built for investigating, not just labeling

Every part of Triage exists to answer one question faster: what actually happened, and why. Here's how it works.

01

Auto-triage

New tickets are automatically summarized, tagged, and routed to the right playbook and team as they arrive — so your queue starts sorted instead of starting from zero.

02

Console

Ask a question in plain language and watch the AI query your databases, pull logs, and check metrics in real time, with every tool call shown as it happens — not a black box.

03

Configurable playbooks

Define playbooks with their own routing hints, triage instructions, and allowed MCP servers per team — a billing playbook and an infra playbook don't need the same access or tone.

04

BYOK

Triage never proxies your model spend through a shared account. Connect your own provider credentials and every call runs against your own contract and rate limits.

05

Read-only by design

Postgres connections run every query inside a read-only transaction, MongoDB's read-only flag is on by default, and the Kubernetes connector explicitly disables every destructive tool. (MySQL's connector doesn't enforce this at the tool level — use a read-only database user if you connect MySQL.) Investigating a ticket should never mean risking the system it's about.

06

Connect your own infra

Triage talks to your infrastructure through MCP servers you configure with your own credentials — no data pipeline, no syncing your database into a third party.

See it work on a real ticket

Free for personal use has every feature unlocked — no seat limits, no watered-down plan.