AI-Powered

Talk to Your Data

Manage data observability through natural conversation. Ask questions, investigate issues, and set up monitoring without leaving your IDE.

Claude Code
You: Is my data healthy?
Claude: Your data health summary:

Overall Status: WARNING

- 45 assets monitored
- 2 tables are stale (orders, customers)
- 1 schema change pending review

Want me to investigate the stale tables?

MCP Tool Calls

The building blocks. Each tool does one specific thing: check freshness, list alerts, create a metric. Your AI agent calls these automatically behind the scenes to get work done.

Think of these like API endpoints your agent can call. You rarely invoke them directly.

Slash Commands

The shortcuts. Each command is a pre-built workflow that chains multiple tool calls together. Type /armor:monitor and the agent figures out which tools to call, in what order, with what parameters.

Think of these like recipes. You say what you want, the agent picks the right tools.

29 Tools

MCP Tool Calls

The primitives your AI agent calls behind the scenes. Every slash command is built on these, and your agent can compose them freely for custom workflows.

Health

Overall system health and asset management

health_summary()list_assets()trigger_asset_discovery()

Freshness

Table freshness monitoring and scheduling

get_freshness_summary()check_freshness()setup_freshness()list_freshness_schedules()

Schema

Schema drift detection and baselines

get_schema_summary()list_schema_changes()create_schema_baseline()enable_schema_monitoring()disable_schema_monitoring()dry_run_schema()

Alerts

Alert management and rule configuration

list_alerts()update_alert()list_alert_rules()create_alert_rule()

Quality

Data quality metrics and validity rules

list_metrics()create_metric()list_validity_rules()create_validity_rule()

Intelligence

AI-powered analysis and natural language queries

ask_question()generate_intelligence()

Catalog

Lineage, tagging, and async job tracking

get_lineage()job_status()list_tags()apply_tags()

Destinations

Alert destinations (Slack, email, webhook)

list_destinations()setup_destination()
15 Skills

Slash Commands

High-level workflows you can invoke by name or just ask naturally

/armor:status

Check data health, alerts, freshness issues, and schema changes across all monitored assets.

Try: "Is my data healthy?"

/armor:alerts

Query and manage alerts with natural language filters. See what fired, resolve issues, and drill into details.

Try: "What alerts fired yesterday?"

/armor:monitor

Set up monitoring for freshness and schema drift. Configure thresholds and alerting in seconds.

Try: "Monitor my orders table for freshness"

/armor:ask

Ask natural language questions about your data. Explore what tables contain, how they relate, and what you know.

Try: "What tables contain customer data?"

/armor:connect

Connect a new data source to AnomalyArmor. Supports Snowflake, Postgres, BigQuery, and more.

Try: "Connect my Snowflake warehouse"

/armor:analyze

Trigger AI analysis for your data assets. Generate intelligence, refresh analysis, and surface insights.

Try: "Analyze my database"

/armor:start

Guided onboarding for new users. Walks you through setup, connecting sources, and first monitors.

Try: "Help me get started"

/armor:investigate

Root cause analysis using lineage, intelligence, and history. Debug stale tables, explain alerts, and trace issues.

Try: "Why is the orders table stale?"

/armor:lineage

Explore data lineage and dependencies. See upstream sources, downstream consumers, and impact analysis.

Try: "What depends on this table?"

/armor:quality

Add data quality checks with metrics and validity rules. Monitor null rates, row counts, uniqueness, and more.

Try: "Add a null check for email column"

/armor:coverage

Analyze monitoring coverage and gaps. Find unmonitored tables, missing alerts, and blind spots.

Try: "What tables have no alerts?"

/armor:profile

Profile tables and columns for statistics and distributions. Inspect cardinality, nulls, and data shape.

Try: "Profile my users table"

/armor:recommend

Get AI-driven recommendations for what to monitor. Suggests tables, thresholds, and coverage priorities.

Try: "What should I monitor?"

/armor:tags

Classify and govern data with tags. Apply PII labels, business domains, and governance categories.

Try: "Tag this table as PII"

/armor:test

Test and preview monitoring configurations before enabling. Dry-run thresholds and preview what would fire.

Try: "Dry-run freshness check on orders"

Works With Your Tools

Integrate with Claude Code, Cursor, or any MCP-compatible AI assistant

Claude Code

Install our plugin for the best experience with slash commands and AI-powered data monitoring.

claude plugin install armor@anomalyarmor

Cursor

Add our MCP server to Cursor for natural language data observability in your IDE.

Add to ~/.config/Cursor/mcp.json

Any MCP Tool

Works with any Model Context Protocol compatible AI assistant.

npx @anomalyarmor/mcp-server setup

Why AI Agents?

Stop context-switching between your IDE and monitoring dashboards. Get answers and take action through natural conversation.

Natural Language

Ask questions in plain English. No need to learn APIs or dashboards.

Instant Answers

Get immediate insights without switching context from your IDE.

Proactive Monitoring

Set up checks and alerts through conversation.

Root Cause Analysis

AI traces dependencies and identifies issues automatically.

Example Workflow

1

"/armor:connect"

Connect your Snowflake warehouse

2

"/armor:analyze"

AI scans your tables and generates intelligence

3

"/armor:monitor"

Set up freshness and schema drift monitoring

4

"/armor:status"

Morning health check, all from your IDE

Ready to Talk to Your Data?

Free tier includes AI agent access. No credit card to start.

Frequently Asked Questions

Ask natural-language questions about your data and get answers in your terminal: "is my data healthy," "what alerts fired today," "why is orders_daily stale," "what should I monitor in this schema." The agent runs real queries through our read-only API and returns grounded answers, not hallucinations.