Skip to content
MCP server

AiSOC

Open-source, self-hostable AI security operations center with auditable agent

Actively maintainedStatus and history

The repository was last updated 23 hours ago and is published under MIT.

Pricing
Open source
Platforms
self-hosted, API, Web
GitHub stars
8.1K
Last commit
1 day ago
Language
Python
License
MIT

Similar MCP servers

About AiSOC

AiSOC is an open-source MCP server for ai models and llm tooling. It is available on self-hosted, API, Web and CLI.

The README describes AiSOC as an MIT-licensed platform that ingests security telemetry, correlates threats into incidents and lets AI agents investigate them, with every prompt and tool call written to an investigation ledger. Proposed responses wait for a human approver before anything reaches a vendor, and a playbook preview mode keeps automated actions off by default.

Detection draws on a library of Sigma, YARA and KQL/EQL-style rules, with per-tenant tuning that changes what the engine fires. The site adds UEBA baselines, signed honeytokens, ATT&CK-mapped purple-team emulation and SOC 2, ISO 27001, NIST CSF and DORA evidence dashboards. The stack runs on Docker Compose or Helm, and the README lists roughly 97 data connectors.

The project suits security teams that need to self-host and review how an AI investigation reached its verdict. The README notes that the default local model is small and that no hosted model provider has been exercised in its testing.

What it does

Auditable agent investigations
Prompts, tool calls, citations and verdicts are recorded in an investigation ledger and can be exported as a signed evidence bundle.
Human-approved response playbooks
Playbooks run in preview by default, and approval steps pause durably until a permitted approver signs off.
Detection rule library
Ships thousands of executable detection rules and translates Splunk, Sentinel and Elastic rules where it can.
Behavioural user analytics
Per-user baselines and Z-score anomaly scoring flag unusual activity for review.
Honeytokens and purple-team tooling
Generates signed deceptive credentials and runs Atomic Red Team and Caldera-based emulation with an ATT&CK coverage heatmap.
Compliance evidence reporting
Dashboards track evidence for SOC 2, ISO 27001, NIST CSF, PCI-DSS, HIPAA and DORA with response-time SLAs.

Good for

  • Triage and investigate security alerts with an AI agent while keeping a reviewable record of its reasoning
  • Run a self-hosted SOC pipeline where telemetry stays inside your own infrastructure
  • Migrate existing Splunk, Sentinel or Elastic detections into a new platform
  • Test detection coverage with purple-team emulation mapped to MITRE ATT&CK
  • Produce compliance evidence for audits under SOC 2, ISO 27001 or DORA

Pricing

Open source

MIT-licensed and self-hostable; no paid plan mentioned in the provided content.

Prices change. Check beenuar.github.io before you buy.

AiSOC FAQ

What license is AiSOC released under?

The README and site state that AiSOC is released under the MIT license.

Can it run without Docker?

The README offers a pip-installed sandbox (aisoc-sandbox) that walks one alert through the agent funnel offline, separate from the full stack.

Does it send prompts to a hosted AI provider by default?

The README says the default install sends no prompt anywhere because the model runs locally beside the stack.

Are response actions automated?

The README says no vendor is touched without a human approver unless a tenant has explicitly granted autonomy for that action.

Is AiSOC free?

Yes. AiSOC is open source, so you can use and self-host it for free.

What are the best alternatives to AiSOC?

The closest alternatives to AiSOC are redamon, Systemores AI Agent & MCP Security Auditor, WhiteBoxXAI, Compliance and Metal Mantra.

Where this comes from

  • Description: written from the project's README and the project's website.

How AIPoweredBot checks servers