AiSOC logo

AiSOC

Self-hosted AI security operations platform with telemetry correlation, agent triage, MITRE ATT&CK investigation, and an auditable MCP interface.

Install AiSOC

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "aisoc": {
      "command": "uvx",
      "args": [
        "aisoc-sandbox"
      ]
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

AiSOC is a self-hosted security operations platform that ingests telemetry, normalizes events, detects threats, groups alerts into incidents, and investigates them with AI agents. Its MCP server makes AiSOC capabilities available to Claude, Cursor, and Continue, while recording agent prompts, tool calls, and decisions for audit. Reach for it when you need an on-premises security workflow with human approval before response actions and no required API keys.

Use cases

  • Correlate security telemetry into incidents
  • Investigate alerts with auditable AI agents
  • Rescan historical events with new threat intelligence
  • Connect Claude, Cursor, or Continue to security operations workflows

Key features

  • Security telemetry ingestion and normalization
  • Alert detection and incident correlation
  • MITRE ATT&CK investigation support
  • Replayable agent decision ledger
  • Human-approved response proposals
  • MCP access for Claude, Cursor, and Continue

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

26/26 checks healthy over the last 16d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 15d ago via OSV.dev · aisoc-sandbox (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars6.7kStargazers on the source repository.
Last commitTodayMost recent push to the default branch.
Directory activity6 viewsConfig copies, upvotes, and views on AllMCPs.

How AiSOC works

What AiSOC does

AiSOC combines security telemetry processing, detection, incident correlation, AI-assisted investigation, and response proposal in one self-hosted system. Incoming events are normalized into a common shape, evaluated against its detection rule library, and grouped into incidents when related activity is identified.

AI agents can investigate incidents using security context and map findings to MITRE ATT&CK. Their prompts, tool calls, reasoning steps, and resulting decisions are recorded in a replayable ledger, giving analysts an auditable trail of how a verdict was reached. Proposed actions remain subject to human approval before they are sent to an external vendor.

The platform also incorporates threat intelligence workflows. New intelligence can be used to rescan previously collected history, which helps identify relevant activity after the original event was stored.

How it works

Telemetry can enter through the ingest API, configured data connectors, or supported collectors. The platform provides vendor-specific normalization for connectors including Splunk, Sentinel, Elastic, CrowdStrike, Okta, AWS, and Kubernetes audit data. A generic mapping is available when a vendor profile is not present.

Cloud and infrastructure collection paths include CloudTrail data delivered through S3 and SQS, GCP Cloud Logging through Pub/Sub, Azure Event Hubs capture, and syslog/CEF. Collectors maintain resumable cursors and limit work per polling cycle so backlogs apply backpressure instead of overwhelming ingestion.

After ingestion, events pass through detection, correlation, and persistence services before becoming alerts available through the API and console. An included MCP server exposes AiSOC to supported AI clients, including Claude, Cursor, and Continue.

Setup and configuration

The standard deployment uses Docker Compose v2. A host needs at least 8 GB of memory and 20 GB of free disk in the Docker VM, along with Python 3.9 or newer and Bash. The first run downloads an approximately 2 GB model into a volume.

Clone the repository and run make up. The startup process creates the environment file, generates service and datastore secrets, bootstraps an administrator, and prints the initial administrator password once. make doctor checks prerequisites, while make smoke sends a test event through ingestion, detection, correlation, and the public API.

Windows and Linux/macOS installation scripts can install prerequisites. A separate aisoc-sandbox package provides an offline demonstration path for walking an alert through the agent funnel without Docker. GitHub Codespaces is also supported as a place to run the Docker-based setup.

Tools and capabilities

  • Ingest security events through the batch API.
  • Connect security products using click-to-connect data connectors.
  • Normalize telemetry and evaluate it against detection rules.
  • Correlate detections into incidents and investigate them with AI agents.
  • Record prompts, tool calls, and decisions in an auditable ledger.
  • Expose AiSOC workflows to Claude, Cursor, and Continue through MCP.
  • Translate Splunk SPL, Sentinel KQL, and Elastic EQL detections when supported by the migration package.
  • Load sample or synthetic data for testing the pipeline.

Limitations and notes

AiSOC is designed for self-hosting and requires a Docker Compose environment with substantial memory and disk. The first startup includes a large model download. Response actions are proposed for human approval rather than executed automatically. Detection migration can refuse rules it cannot carry accurately, and translated rules may be partial.

The sample and demo datasets are synthetic and should not be treated as customer incidents or benchmark results. The MCP client list explicitly includes Claude, Cursor, and Continue; MCP setup should be evaluated alongside the deployment and authentication configuration used for the AiSOC instance.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about AiSOC

AiSOC is an MCP server that connects Claude, Cursor, and Continue to a self-hosted security operations platform. Its main capabilities cover security telemetry, alert and incident workflows, AI-assisted investigation, MITRE ATT&CK analysis, and auditable agent decisions, with human approval required before proposed responses reach vendors.

Install AiSOC by cloning the repository and running `make up` with Docker Compose v2. The setup creates the environment file and required secrets, then starts the self-hosted console and services. The deployment requires Python 3.9+, Bash, 8 GB of memory, and 20 GB of free Docker disk.

Yes, AiSOC ships an MCP server for Claude, Cursor, and Continue. These clients can connect to the self-hosted AiSOC platform through its MCP interface and use the security operations capabilities exposed by the deployment.

Yes, AiSOC accepts events through its ingest API and supports configured data connectors. Vendor-specific coverage includes Splunk, Sentinel, Elastic, CrowdStrike, Okta, AWS, and Kubernetes audit data, while generic mapping supports sources without a dedicated vendor profile.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewAiSOC AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/aisoc?style=directory)](https://allmcps.com/mcp/aisoc)
HTML Embed
<a href="https://allmcps.com/mcp/aisoc"><img src="https://allmcps.com/api/badge/aisoc?style=directory" alt="AiSOC on AllMCPs" /></a>