Ask Zabbix what is alerting now, which hosts are affected and what an item measured.
Connect Zabbix with your frontend's host and an API token, and Memtro reads current problems with severity, hosts and acknowledgement, event history, hosts, maintenance windows and items with their recent values. The hourly sync writes each new problem event into memory. When you ask, and once an administrator has granted them by name, it can comment on, acknowledge, re-grade, suppress, rank or close problems, set maintenance windows, turn monitoring of hosts, triggers or items off or on, and add hosts. Maintenance and disabling silence alerts, closing a problem or deleting a host is permanent, and configured Actions may notify people.
Ask in plain English from chat, your editor, a scheduled job or your own agent. The model picks the right Zabbix tool, reads what it needs, and answers with the source cited.
These tools change something in Zabbix. The model is told to use them only when you ask for exactly that, and an administrator grants each one by name to the teams that may use it; a personal connection acts only after its owner allows it.
Once Zabbix is connected, these are the kinds of things people ask. Plain questions, no syntax. If the answer lives partly in another system you have connected, Memtro looks there too.
Zabbix + MemtroWhich high-severity problems are unacknowledged right now?
What alerted on db-prod-02 overnight?
What was the CPU load on web-01 over the last 24 hours?
Which hosts are in maintenance at the moment?
Put db-prod-02 into maintenance from 22:00 to midnight tonight
The on-call engineer at Northwind asks what happened overnight. Memtro reads Zabbix events for the last twelve hours and lists four problems, three resolved, and one disk space warning on db-prod-02 still open and unacknowledged since 03:40.
Priya Nair asks Memtro to acknowledge the db-prod-02 disk problem with a note that she is extending the volume. Memtro acknowledges it as the token's user, and any Zabbix Action set up for updates may notify the team.
The teams that lean on Zabbix most, a question each of them asks, and what they connect beside it.
A record in Zabbix is rarely the whole story. Connect the systems that hold the rest and Memtro checks them together, then remembers what it learned so the next question starts further along.
Using something else for the same job, or both? Each connector has its own tools; connect as many as you like.
UptimeRobot
Honeycomb
Cronitor
FireHydrant
StatusCake
GlitchTip
GLPI
Langfuse
LibreNMS
NetBox
Snipe-IT
Woodpecker CI
LangSmith
Helicone
GrowthBook
Drata
Intruder
GitGuardian
Secureframe
ilert
JumpCloud
ConfigCat
Split
Aikido Security
Astronomer
Comet Opik
CrowdStrike Falcon
Detectify
Domotz
LambdaTest
LangWatch
LlamaCloud
Logz.io
Lunary
NinjaOne
Northflank
Portkey
PromptLayer
PRTG Network Monitor
Raygun
Retell AI
Runpod
Shodan
Spike.sh
Squadcast
Tenable
UpGuard
Uptime.com
Vanta
Zenduty
Cortex
WorkOS
Kandji
Automox
Huntress
SimpleMDM
Healthchecks.io
Rapid7 InsightVM
ZenhubNeeds Zabbix 5.4 or later (7.0 or later recommended), and a dedicated user whose role allows API access, with read access to the host groups wanted and read-write only where problem actions should work. Maintenance, host, trigger and item changes need an Admin or Super admin user type with read-write on those host groups
In Zabbix open Users → API tokens → Create API token, name it Memtro, choose that user, add, and copy the token: it is shown once
In Connections → Zabbix in Memtro type your frontend's name in Host only, with no https://, path or port, and paste the API token. The host must be a public DNS name with a publicly trusted certificate
Ask which high-severity problems are unacknowledged right now
Chat in the dashboard, the OpenAI-compatible API, scheduled jobs, and any MCP client: Claude Code, OpenCode, Cursor, Claude Desktop. The Zabbix tools are the same everywhere.
Only when you ask, and only what it has been granted: add a comment to a problem; acknowledge a problem; unacknowledge a problem; change a problem's severity; suppress a problem, until a time or indefinitely; unsuppress a problem; close a problem by hand; mark a problem as the cause of others; mark a problem as a symptom of another; put hosts or host groups into a maintenance window (suppresses their alerts); change a maintenance window's times, hosts or groups; delete a maintenance window (permanent); turn monitoring of a host on or off; add a host to monitoring; delete a host and its history (permanent); enable or disable a trigger; enable or disable an item. Every other tool reads, searches or lists. The Connections page labels the tools that act, an administrator grants each one by name to the teams that may use it, and a personal connection acts only after its owner allows it.
A personal connection serves only your own requests. A shared connection serves the organisation: everyone, or only the teams an admin chooses. Admins can always add shared connections; other members need the permission.
Credentials are encrypted at rest. Tool results are cached briefly so agent loops do not repeat identical calls, and the facts the model extracts are saved to your organisation's memory with provenance (system, record and link) so they can be checked and superseded. An hourly sync anchors recently changed Zabbix records into memory as entities.
Host and API token. Needs Zabbix 5.4 or later (7.0 or later recommended), and a dedicated user whose role allows API access, with read access to the host groups wanted and read-write only where problem actions should work. Maintenance, host, trigger and item changes need an Admin or Super admin user type with read-write on those host groups.
Seven-day free trial, then one monthly price per organisation. Connectors, jobs and memory are never metered.
Sign in, add a provider key, create an API key and point your tools at Memtro. It takes about five minutes.
7-day free trial · card on file, nothing charged until it ends · cancel any time
