AI-Powered Configuration Intelligence for Enterprise Networks
Ask your network in plain English, or script it in Python.
Iris, the AI centre in rConfig, reads your live configurations, backup history and compliance baselines. Ask it a question in plain English, or have it write a Python script against the rConfig API. Iris is read-only to the network. It never changes a device.
- 10 AI capabilities
- 7 read-only MCP tools
- 13 AI providers
- 0 per-prompt fees
- Read-only to the network
$ iris ask “Audit all Cisco devices for NIST 800-53 deviations and write remediation scripts”
- parse query, scope and intent ok (84 Cisco devices)
- fetch running configs via MCP ok (84 configs)
- analyse against NIST 800-53 ok (12 violations, 7 devices)
- write Python remediation scripts ok (7 scripts)
- compile compliance report ok (ready)
- MCP tools
- 7
- Providers
- 13
- Device changes
- 0
Iris works inside your network configuration manager
Live topology
Answers take your live topology into account.
Compliance
Drift detection against your compliance baselines.
Inventory
Asset intelligence from your device inventory.
History
Change tracking across your backup history.
Two ways to use Iris
GenAI: plain-English analysis
Ask questions about configurations and security posture.
- Find routing loops and misconfigurations.
- Explain a live config in plain English.
- Detect compliance drift against NIST and CIS.
- Use any of 13 AI providers, in the cloud or inside your own network.
MCP: automation for developers
Combine Python automation with the Model Context Protocol.
- Write ready-to-run backup and audit scripts.
- Explore the rConfig APIs by asking.
- Write Python, Bash and PowerShell workflows.
- Follows your RBAC roles and writes audit logs.
| Feature | GenAI: plain-English analysis | MCP: automation for developers |
|---|---|---|
| AI type | Generative (natural language) | Model Context Protocol (developer) |
| Core use | Troubleshooting, config analysis | Automation, scripting, integration |
| Primary users | Network engineers, security analysts | Developers, DevOps, MSP teams |
| Integrations | 13 AI providers, cloud and local | Claude Code, Gemini Code Assist |
| Languages | Natural-language queries | Python, PHP, Bash, PowerShell |
| Security | Drift detection, ISO and NIST context | RBAC and audit logging |
| Licensing | Included free | Included free |
You choose the model and where it runs
GenAI and MCP in one platform
Conversational insight and developer automation sit in the same network configuration manager.
Python automation ready
Write Python scripts wired to the rConfig API, then schedule them with cron.
Runs inside your security model
Iris follows your RBAC roles, access logs and change control. With a cloud provider, the connection goes directly from your server to that provider.
Open, with your own keys
No lock-in and no per-prompt fees. Use your existing AI provider credentials.
13 AI providers
Cloud (11)
- Anthropic
- OpenAI
- Azure OpenAI
- AWS Bedrock
- Google Gemini
- Mistral
- Cohere
- DeepSeek
- Groq
- xAI
- OpenRouter
Inside your own network (2)
- Ollama
- OpenAI-compatible
How teams use Iris
Troubleshooting with AI
A routing table keeps flapping. GenAI reads the live configs and state history and points to the likely cause. It can suggest commands for you to review.
Learn moreAutomating with Python
To compare VLANs across 20 data centres, MCP writes a Python script against the rConfig API that fetches, compares and reports.
Learn moreCompliance and auditing
GenAI checks your ACL posture from a plain-English question. MCP writes audit scripts aligned to ISO and NIST standards.
Learn more
Feature deck
Iris feature deck
The Iris feature overview as a deck (version 1.2), for walking your team through what Iris does.
Open the feature deckOne-pager
Iris one-pager
Iris on a single page (version 3.0), ready to share with the people who sign off the decision.
Open the one-pager
AI should empower engineers, not bill them monthly or harvest their data. We built rConfig AI to be open, private, and incredibly useful, whether you're debugging a router or designing automation in Python.
The thinking behind Iris
The network AI maturity model
Four staged levels of network AI mapped to the recognised TM Forum autonomy scale, and where rConfig sits today.
The Iris approach
Private by design, grounded in real network knowledge, and clear about what it does today. You bring the model, and your configs stay in your network.
Iris FAQ
What is Iris and how does it help with network automation?
Iris is the AI centre in rConfig, a set of AI tools built into rConfig V8. It combines Generative AI, which analyses your configurations in plain English, with Model Context Protocol (MCP) AI, which helps developers write automation scripts (Python, Bash and others) faster.
Why does Iris have both Generative AI and MCP?
Engineers need two kinds of help: understanding complex data (GenAI) and building automation tooling (MCP). Both sit inside the network configuration manager, so what you learn from one feeds the other.
How does Iris fit into Python-based network automation workflows?
MCP turns what you ask for into code. Ask it to 'Write a script to back up Cisco devices in Group A' and it writes Python that uses the rConfig API libraries, for you to review.
What makes GenAI in Iris different from generic AI assistants?
Context. A generic assistant does not know your network. GenAI in Iris has access to your device inventory, live configurations and compliance reports, so its advice refers to your own devices.
Does Iris require internet access or external cloud processing?
The rConfig platform runs on-premise. For AI features, you bring your own API keys (OpenAI, Anthropic and others), or run a model inside your own network with Ollama or an OpenAI-compatible server. With a cloud provider, the connection goes directly from your server to that provider, and rConfig (the company) never sees your data.
Can Iris be used in large enterprise or MSP environments?
Yes. Iris follows the platform's standard RBAC (role-based access control), so you can limit which users have access to AI features. All AI-generated actions are logged for audit.
See Iris on your own configs
Iris works from your configurations and backup history. Request a demo and ask it a question about your network.

