logo

Resolve Linux and FreeBSD incidents faster

with AI that stays under your control.

Troubleshoot interactively, run guarded automation, and turn alerts into pre-analyzed tickets.

Run your first AI-assisted server diagnosis in minutes.

Start free trial

Choose your workflow
↙
↓
↘
Troubleshoot incidents live with AI
ai - interactive co-admin
  • Every command requires explicit confirmation (Human-in-the-loop).
  • Uses real command output, logs, and system context.
  • Investigates before proposing changes.

Watch nginx 502 diagnosis
Learn More
Run guarded automation workflows
ac-ops - policy-controlled execution
  • Define which tools, targets, and actions are allowed.
  • Keeps runs predictable with timeouts and output limits.
  • Return structured JSON or text for scripts, CI, or tickets.

Watch guarded automation demo
Learn More
Route alerts into pre-analyzed tickets
ac-gateway - event-driven workflows
  • Receive and process alerts from monitoring systems via webhooks.
  • Map alerts to hosts, policies and use cases for analysis.
  • Send enriched summaries to ServiceNow, email, Slack, ...

See alert-to-ticket workflow

Debian Ubuntu RHEL Fedora FreeBSD and more

Built for server operations and DevOps: clarity, traceability, and control.

For MSPs, server fleets, and alert-driven workflows, book a technical demo.

Book technical demo

Reviewed by technical press

Independent hands-on coverage of Admin Companion.



“The direct integration of Ayonik Admin Companion into the Linux console quickly proved to be an invaluable help in practice.”


IT-Administrator

“System commands and the AI itself are well isolated. Nothing happens without the system user’s consent.”


IT-Administrator

Logos are trademarks of their respective owners and are used for reference purposes only.

Designed for production environments

Admin Companion is built to reduce operational risk while speeding up incident response.

Curated admin references
Higher precision in Linux / FreeBSD ops than generic chat due to usage of curated admin documentation.
No unattended commands
Admin Companion proposes commands and explains the rationale. You approve execution step by step in interactive mode.
Built for real shell context
It works from command output, logs, OS facts, and task context - not isolated chat prompts.
Policy controlled automation
For repeatable workflows, restrict allowed tools, targets, timeouts, and outputs.

ai - interactive co-admin

Troubleshoot safely, stay in control

  • Human-in-the-loop: every command requires explicit confirmation.
  • Investigate-first workflow before proposing changes.
  • Uses real command output step-by-step.
Watch interactive sessions Download & Install

What you get

ai - interactive co-admin

Less time firefighting. Faster recovery. Safer changes.
timer

Faster first diagnosis

Ask Admin Companion to inspect a failing service, recent logs, ports, configs, and dependencies. It proposes the next safe command and explains why.

security

Safer changes

All actions are visible before execution. Nothing runs without approval in interactive mode.

library_books

Fewer context switches

Work directly in the terminal instead of copying logs into generic chat tools.

terminal

Better answers for Linux / FreeBSD operations

OS-aware and topic-aware references help reduce irrelevant advice.

How it works

ai - interactive co-admin

Describe → propose & explain → confirm & execute.
chat

1) Describe the task

Use natural language directly in your shell.

ai "Investigate and fix 502 error in nginx"

lightbulb

2) Get a plan + next commands

Admin Companion proposes the next commands and explains why. You review the plan before anything executes.

verified_user

3) Confirm execution

Nothing runs unattended. You approve each command before it executes. Output from approved commands is used for the next step.

You stay fully in control.

Why it stays accurate and relevant

ai - interactive co-admin

Curated admin references + OS, topic and local system context aware.

Curated knowledge base

Admin Companion uses curated Linux/FreeBSD administration references and selects the right knowledge for your OS and topic to reduce irrelevant or unsafe advice.

OS-aware: selects the right context for your distribution.

Debian Ubuntu RHEL Fedora FreeBSD and more

Topic-aware: activates domain knowledge when relevant.

Kubernetes Ansible nginx Apache HTTP server MariaDB Postfix

4-layer local context system

A multi-layer context system keeps troubleshooting consistent across sessions: session, task, environment facts, and durable learnings.

1) Dialogue history
(session context)
Automatically managed for coherent multi-step troubleshooting.

2) Topic context
(task-scoped)
User-managed task focus across many steps.

3) Background context
(environment facts)
User-managed stable facts (paths, service names, distro conventions, policies).

4) Long-term context
(persisted learnings)
Automatically managed durable notes you can clear any time.

Examples:

Background: "logs shipped to journald only, not in /var/log" or "app lives in /opt/acme/app/current"

Topic: "This incident is about: 502 error on nebula-three after last deploy."

All context information is stored locally and can be inspected and cleared at any time.

Watch real terminal sessions

ai - interactive co-admin

Production-style workflows: analyze → propose → confirm → execute.

Real terminal incident sessions

Scenario: 502 Bad Gateway across NGINX → Apache → Python

Admin Companion analyzes logs and service state, proposes the next command, and waits for confirmation before execution.

bad
↓

Command:
ai "Our nginx server returns an error on port 5000. Please investigate and fix."

↓

good
Start free trial

Scenario — Clean up disk space safely

Analyze disk usage, propose safe cleanup steps, and execute only after confirmation.

Command: ai "Check disk usage and suggest cleanup"

Scenario — Generate automated backup scripts

Generate a documented shell script tailored to your environment that you can review and adapt.

Scenario — Find and install missing packages

Identify the correct package for a missing tool and propose the safe installation steps.

ac-ops - policy-controlled automation

Run repeatable operational checks and ticket-generation workflows with explicit tool allowlists, target restrictions, timeouts, and bounded output.

Built-in guardrails

Only approved tools can be called. Per-tool restrictions, and optional audit logs keep automation predictable and tracable.

Bounded automation runs

Timeouts and output limits for predictable runs. Writes primary output to stdout in JSON or text.

What it is for

Use ac-ops for repeatable diagnostic workflows: scheduled checks, runbook steps, incident enrichment, and ticket generation.

Typical use-cases: Issue analysis for: Docker, Kubernetes, Terraform, Ansible

Documentation: ac-ops

Policy-controlled automation demo

ac-ops - guard-railed automation

Scenario: Docker error analysis with automated ticket creation

What happens here

  • A real Docker 'die' event is piped into ac-ops
  • ac-ops analyzes the issue with restricted read-only tools
  • It generates a ticket with findings and recommended next steps
Why it is safe
  • Policy-controlled tool access
  • Bounded execution
  • No unrestricted automation
Note: Gray text = demo annotation. Highlighted orange output = generated ticket.

Guard-railed ops automation designed for production workflows.

ac-gateway - event-driven workflows

Route monitoring alerts to the right server and analysis workflow.

Receives webhook alerts, matches them to profiles, target hosts, and use cases, then triggers controlled ac-ops analysis and forwards the result.

What it is for

  • Accepts webhook events such as Prometheus Alertmanager alerts.
  • Maps payloads to target hosts and use cases.
  • Triggers remote ac-ops analysis without glue code.
ac-gateway Documentation

Built-in guardrails

  • Bearer-token protected HTTP endpoints.
  • SSH to target hosts with least privilege.
  • Allow/deny target policies and use-case restrictions.
ac-gateway Documentation

Route alerts into controlled analysis and ticket-ready summaries - without custom glue code.



Sample: Prometheus → ac-gateway → ac-ops → ServiceNow / Email / Slack ticket

For MSPs and IT service providers

Operate across many customer servers from one account with controllable API keys.

One account for different customer environments

Create multiple API keys and use them across different servers or customers.

Impact: simple separation of access by customer or purpose.

Key lifecycle control

Activate or deactivate individual API keys at any time without affecting others.

Impact: quick offboarding if a key is exposed.

Talk to us about MSP workflows

Ready to reduce ops time without losing control?

Install the client, get free credits, and try it on a real incident or routine task.
Download & Install Create account and get free credits Contact Sales

Ideal for teams who need speed, precision, and strong operational control.