Akmatori

Free plan Open source APILinuxMacSelf-hostedWebWindows

What Akmatori is

Akmatori turns production alerts into AI-assisted incident response workflows for SRE, platform, and DevOps teams. Its agents gather logs, metrics, deployment details, and runbook knowledge to investigate incidents and suggest next actions. It accepts alerts from Alertmanager, PagerDuty, Grafana, Datadog, and Zabbix, and can monitor Slack channels. Agents can reach infrastructure through SSH, Kubernetes, scripts, and API clients; risky proposed actions require approval in the web UI or Slack. The project is open source under the Apache License 2.0, with a free plan at no cost. It can be deployed with Docker Compose or Helm, including on-premises with local models for air-gapped setups. It supports multiple LLM providers and recurring investigations that can post results to Slack. The maker says incident data stays in the customer’s infrastructure unless an external provider is configured. Akmatori also documents a REST API with incident operations, health checks, authentication, webhooks, and an OpenAPI 3.0 specification.

Who it suits

Akmatori suits SRE, platform, and DevOps teams that want to investigate production alerts with AI-assisted workflows. It may fit teams seeking self-hosted deployment, Kubernetes access, and approval controls for risky actions.

What it does well

Investigates alerts using logs, metrics, and runbook knowledge.
Accepts alerts from five named monitoring services.
Risky actions require approval before execution.
Supports on-premises deployment and local models.
Open source under Apache License 2.0.

What to know first

Risky proposed actions require a human approval.
External providers may receive incident data when configured.
Bug reports and product questions go through GitHub issues.

Plans

Plan	Billed and limits	Price
Open-source	Apache 2.0 · Self-hosted deployment	Free
Vendor pricing	Pricing supplied by vendor on request · Capterra lists no free trial · Capterra lists no free version	On request

As akmatori.com lists them, read 7 Oct 2026. The maker's own page has the final word.

Getting started

Open https://akmatori.com/.
Choose Docker Compose or Helm for deployment.
Configure alert sources such as Alertmanager, PagerDuty, Grafana, Datadog, or Zabbix.
Connect a supported model provider or an OpenAI-compatible endpoint.
Set up agent access to infrastructure through SSH, Kubernetes, scripts, or API clients.

Verdict

Akmatori combines alert investigation, infrastructure access, and guarded remediation in an open-source package. Teams should review its approval workflow and data handling, especially if configuring external providers.

Questions

How much does Akmatori cost?

The Open-source plan is free under the Apache License 2.0.

Is Akmatori open source?

Yes. The project is released under the Apache License 2.0 and developed in the open.

Where can Akmatori be deployed?

It can be deployed with Docker Compose or Helm, including fully on-premises with local models for air-gapped setups.

Which alert sources does Akmatori accept?

It accepts alerts from Alertmanager, PagerDuty, Grafana, Datadog, and Zabbix, and can monitor Slack channels.

Which model providers does it support?

It supports OpenAI, Anthropic, Google, OpenRouter, NVIDIA NIM, MiniMax, and OpenAI-compatible endpoints, including on-premises models.

How are risky actions handled?

Proposed risky actions require approval in the web UI or Slack before execution.

Facts from Akmatori's own pages

Agent investigations: Each incident runs a full agent session that gathers logs, metrics and context, with subagents for runbook lookup and cross-incident memory. · github.com, 15 Oct 2026
Alert ingestion: It accepts alerts from Prometheus Alertmanager, PagerDuty, Grafana Alerting, Datadog, Zabbix, Slack channels and scheduled cron jobs. · akmatori.com, 15 Oct 2026
Alert integrations: It accepts alerts from Alertmanager, PagerDuty, Grafana, Datadog, and Zabbix, and can monitor Slack channels. · akmatori.com, 7 Oct 2026
API: The REST API documents incident operations, health checks, JWT and API key authentication, webhooks, and an OpenAPI 3.0 specification. · akmatori.com, 7 Oct 2026
API authentication: Most API requests require JWT or API-key authentication, with X-API-Key, Authorization: ApiKey and Bearer formats accepted. · akmatori.com, 15 Oct 2026
Approval workflows: Risky remediation actions become proposals that an operator can approve in the web interface or Slack before execution. · akmatori.com, 15 Oct 2026
Architecture: The stack uses API server, PostgreSQL, agent worker, MCP gateway and web UI containers, with network isolation and no database access for the agent worker. · github.com, 15 Oct 2026
Context files: Reference documents can be uploaded as context files for agent incident analysis. · github.com, 15 Oct 2026
Credential handling: Tool credentials are encrypted at rest, resolved by the MCP gateway at execution time and kept away from the AI agent; executions are logged for audit. · akmatori.com, 15 Oct 2026
Data control: The maker says incident data stays in the customer’s infrastructure unless an external provider is configured. · akmatori.com, 7 Oct 2026
Data residency: Self-hosted deployments keep incident data inside the operator's infrastructure, and local-model mode avoids external AI egress. · akmatori.com, 15 Oct 2026
Deployment: Akmatori can be deployed with Docker Compose or Helm, including fully on-premises with local models for air-gapped setups. · akmatori.com, 7 Oct 2026
GitHub activity: The main public repository shows 42 stars, 8 forks and 1,046 commits on the reviewed page. · github.com, 15 Oct 2026
Infrastructure tools: Agents can access infrastructure through SSH, Kubernetes, scripts, and API clients. · akmatori.com, 7 Oct 2026
Investigation: Its agents gather logs, metrics, deployment context, and runbook knowledge to investigate incidents and recommend next actions. · akmatori.com, 7 Oct 2026
License: The source code is licensed under the Apache License 2.0. · github.com, 15 Oct 2026
Limit: independent feedback: GetApp reports zero user reviews and assigns zero feature ratings for data aggregation, IT monitoring and root-cause analysis. · getapp.com, 15 Oct 2026
Limit: pricing transparency: Capterra says buyers must contact the vendor for pricing rather than showing a public price. · capterra.com, 15 Oct 2026
Limit: trial availability: Capterra lists no free trial and no included free version. · capterra.com, 15 Oct 2026
LLM providers: Supported providers include OpenAI, Anthropic, Google, OpenRouter, NVIDIA NIM, MiniMax, Ant Ling and custom OpenAI-compatible endpoints. · akmatori.com, 15 Oct 2026
Monitoring tools: Built-in tool types include SSH, Zabbix, VictoriaMetrics, Grafana, PagerDuty, Kubernetes and declarative HTTP connectors. · akmatori.com, 15 Oct 2026
On-prem models: The documentation supports on-prem or air-gapped operation with local OpenAI-compatible models such as Ollama, vLLM, GLM, Kimi, Mistral and LLaMA. · akmatori.com, 15 Oct 2026
OpenAPI and SDKs: Every installation serves interactive Swagger documentation and an OpenAPI 3.1 YAML contract; the reviewed documentation does not list official language SDKs. · github.com, 15 Oct 2026
Organization details: The Akmatori GitHub organization identifies itself as AI-Powered Incident Management, lists the United States of America and shows two public Go repositories. · github.com, 15 Oct 2026

Where Akmatori stands on Freedom251

AI DevOps Assistants#1 of 26