Multi-Domain AI Platform

AI agents that work with your data, your tools

AgentsIQ delivers intelligent agents for infrastructure and operations teams. Each agent is a standalone plugin — deploy in minutes, configure through the UI, no complex setup.

3
Live agents today
2
Env vars to deploy
<15m
From clone to running
AgentsIQ Studio
Alert Analyser
CUR Analyser
Infra Health
G
What is my alert noise rate this week?
Your noise rate is 66.7% across 6 alerts. 4 are noise (P3/P4, auto-close <5min), 2 are genuine — P1 database failure and P2 Redis failover, both still open and unacknowledged.
G
Which alerts should I suppress?
Suppress cpu-high, disk-space-warning, ssl-cert-expiry, and memory-pressure — all auto-close in under 5 minutes with 100% auto-resolve rate. Zero signal value.
How it works

One platform.
Unlimited agents.

Every agent is a standalone plugin with its own repo and configuration. Deploy one or twenty — the platform handles the rest.

🔌

Standalone Plugins

Each agent is a self-contained repository. Clone it, set two environment variables, and run docker compose up. No platform SDK, no lock-in.

⚙️

UI-Configured

Every agent exposes a Settings page. Credentials, data sources, thresholds, and sync intervals — all configured through the browser, stored encrypted in your database.

🗄️

Persistent by Default

Configuration and data survive restarts. Agents reconnect to your data sources automatically on startup. No manual reconfiguration after upgrades.

📋

Manifest-Driven

Each agent declares its capabilities in a manifest.json. The platform reads it and auto-generates the Settings UI — new agents get a config page for free.

Register in seconds

Deploy your agent anywhere — Railway, EKS, ECS, or bare metal. Register it in the portal with a name, slug, and invoke URL. It appears in the catalogue instantly, ready for chat.

# Register a new agent POST /api/registry/agents { "name": "Kafka Analyser", "slug": "kafka-analyser", "invoke_url": "http://kafka:8080" } → Appears in catalogue immediately
Available today

Three agents.
Real problems solved.

Production-ready agents that connect to your existing tools — no new infrastructure required.

Alert Intelligence

Alert Analyser

Connects to OpsGenie or JSM and separates genuine incidents from automated noise. Identifies repeat offenders and tells you exactly what to suppress.

  • Live OpsGenie / JSM API integration
  • Noise scoring with tunable thresholds
  • Suppression recommendations
  • Scheduled auto-sync
Cost Intelligence

CUR Analyser

Upload your AWS Cost and Usage Report and ask plain-English questions about your spend. Get breakdowns, trends, and saving recommendations instantly.

  • CSV file upload or synthetic data
  • Cost breakdown by service and region
  • Anomaly detection
  • AWS Cost Explorer API (Phase 2)
Infrastructure

Infra Health Check

Monitors service health, uptime, and incident status across your infrastructure. Ask about degraded services and get remediation recommendations.

  • Service health monitoring
  • Uptime and response time tracking
  • Incident history
  • Live API integration (Phase 2)
Coming Soon
Streaming

Kafka Analyser

Consumer lag, broker health, topic throughput, connector status — ask anything about your Kafka cluster in plain English.

Getting started

From zero to running
in under 15 minutes.

Two environment variables. One docker compose command. Everything else configured through the browser.

1

Clone the agent repo

Each agent has its own public repository under the AgentsIQ GitHub organisation. Clone only what you need.

2

Set two environment variables

DATABASE_URL and ENCRYPTION_KEY. That's it. API keys, data sources, and thresholds are all configured through the UI.

3

Run docker compose up

The agent starts, creates its database tables, and is ready to accept requests. No migrations to run manually.

4

Configure through the browser

Open the Settings page, enter your data source credentials, click Save & Sync. The agent is ready for questions.

5

Deploy to your platform

The k8s/ folder includes Kubernetes manifests and a deploy.sh script for EKS or any Kubernetes cluster. HPA included.

bash
# 1. Clone
git clone github.com/agentsiq/alert-analyser
cd alert-analyser

# 2. Two env vars only
cp .env.example .env
DATABASE_URL=postgresql+asyncpg://...
ENCRYPTION_KEY=your-generated-key

# 3. Start
docker compose up

# 4. Verify
curl http://localhost:8080/health
→ {"status": "ok"}

# 5. Ask your first question
curl -X POST localhost:8080/invoke \
  -d '{"user_message": "What is my noise rate?",
      "session_id": "abc-123",
      "context": {}, "history": []}'

# Kubernetes deployment
cd k8s && ./deploy.sh \
  --image registry/alert-analyser:v1.1.0 \
  --namespace ai-agents \
  --db-url "postgresql+asyncpg://..."

Ready to deploy
your first agent?

Try the live demo, browse the source on GitHub, or get in touch to discuss deploying AgentsIQ in your organisation.