Portal MCP Server

Overview

The Nasuni Portal MCP Server enables AI-powered management and monitoring of your Nasuni infrastructure. By connecting supported AI clients, such as Claude, ChatGPT, and Microsoft Copilot, directly to your Nasuni Portal, you can retrieve insights, generate reports, and diagnose issues using natural language.

The server exposes a comprehensive set of Ops IQ tools covering appliance health, performance telemetry, data protection, fleet-wide TCO analysis, and more — all accessible through a secure, OAuth 2.0-authenticated connection.

Note: The Portal MCP Server is currently in private preview. Contact your account manager to participate in the private preview

What is MCP?

The Model Context Protocol (MCP) is an open standard that allows LLM Clients and AI assistants to connect to external data sources and tools in a secure, standardized way. Think of it as a universal adapter between AI clients and the systems they need to interact with.

With MCP:

AI clients discover available tools automatically — no manual API integration required
Authentication is handled via OAuth 2.0, ensuring secure access
Tools are described in a machine-readable format, so the AI understands what each tool does, what inputs it needs, and what it returns
You interact with your infrastructure using natural language instead of navigating dashboards or writing scripts

The Portal MCP Server implements this protocol, turning your Nasuni Portal into an AI-accessible data source.

What is the Portal MCP Server?

The Portal MCP Server is a hosted service that bridges your Nasuni Portal with AI clients. It provides:

Tools — A suite of tools for observability, diagnosis, and optimization of Nasuni Edge appliances, volumes, and cloud infrastructure
OAuth 2.0 Authentication — Secure sign-in using your existing Nasuni Portal credentials, with role-based access control (RBAC) inherited from your Portal account
Multi-Client Support — Works with Claude (Desktop and Web), ChatGPT (Web and Desktop), and Microsoft Copilot Studio
Regional Deployment — MCP server endpoints are available per region to minimize latency

What You Can Do

Ask your AI assistant questions like:

“Are any latest updates available to appliance demoEdge?“
“Give me a list of all my volumes, used capacity, and compression savings.
“Look at activity logs and summarize recent changes in Portal. “
“List all the appliances connected to volume demoVol.”

With Ops IQ tooling, you can do more:

"What's the health status of all my Edge appliances?"
"Show me the data protection metrics for Volume3 over the last 7 days."
"Are there any appliances with high cache disk latency?"
"Compare snapshot protection timing across my US volumes."
"What would I save by rightsizing my fleet?"
"Show me activity logs from the last 24 hours."

The AI client calls the appropriate Ops IQ tools, retrieves data from your Portal, and presents the results in a conversational format — no dashboards, no queries, no context-switching.

Client Setup Guides

Supported Clients

The Portal MCP Server should be supported by any LLM Client supporting MCP. The following clients have been tested.

Claude Desktop and Web
ChatGPT Desktop and Web- Requires developer mode
Microsoft 365 Copilot Studio

Prerequisites

Before connecting any AI client to the Portal MCP Server, ensure you have:

A Nasuni Portal account with MCP Server access (currently in private preview)
Portal credentials — you will authenticate via OAuth 2.0 using your existing Nasuni login (native user or SSO credentials)
Your MCP Server URL — retrieved from the Portal UI (see below)
1. Log in to the Nasuni Portal UI.
2. Navigate to Account Settings → Integrations → Portal MCP Server.
3. Copy the MCP Server URL displayed on the page.
Development/Developer Mode enabled in your AI client (required for ChatGPT)

Claude (Desktop and Web)

The setup process is identical for both Claude Desktop and Claude Web.

Prerequisites: Development Mode must be enabled in your Claude account settings.

Steps:

Go to Account → Settings → Connectors and click Add custom connector.
Enter a name for the connector (e.g., Nasuni Portal) and paste your MCP server URL into the URL field.
Click Add, then click Connect on the confirmation screen.
Complete the OAuth authentication flow by signing in with your Nasuni credentials. Once authenticated, you will be redirected back to the Claude client.

The Nasuni MCP connector is now active and ready to use in your Claude conversations.

ChatGPT (Web and Desktop)

Important: Custom connector setup in ChatGPT requires Developer Mode and is available on the web version only. Once you add the custom connector via the ChatGPT web, it will appear in the Desktop version automatically.

Requirements:

Web version of ChatGPT only (chatgpt.com)
A paid plan: Plus, Pro, Business, Enterprise, or Edu
For Business or Enterprise workspaces, your administrator may need to enable custom connectors before this option becomes available

Steps:

Open chatgpt.com and enable Developer Mode from your account settings.
In the left sidebar or settings menu, locate the Connectors section.
Click Add connector and select the option to add a custom MCP server.
Enter your MCP server Name and URL. Set Authentication to OAuth and Registration method to Dynamic Client Registration (DCR).
Hit Create. This will take you to authenticate using your Nasuni credentials, then redirect you back to the client.
Scroll down to see all available tools listed.

Microsoft Copilot Studio

This walkthrough assumes you have already created an agent in Copilot Studio. If you have not, please follow the Copilot Studio quick start guide.

Steps:

Start on your Copilot Studio agent overview page. Click on the Tools tab.
On the top left of the page, click the Add a tool button. An overlay will appear showing available tools.
On the overlay, click the Model Context Protocol tile.
Enter your Portal MCP server URL. For Authentication, select OAuth 2.0. For the "Type" section, select Dynamic discovery.
Click Create. You will be directed through the OAuth process.
On the top right of the page, click Settings, then select Connection Settings on the left nav bar. Connect the new MCP server.
Click the Test button at the top right of the page. Ask your agent to list the tools now available from the newly connected server.

Authentication & Security

OAuth 2.0

The Portal MCP Server uses OAuth 2.0 for authentication. When you connect an AI client for the first time, you are redirected to the Nasuni Portal login page to sign in with your existing credentials. No API keys or service accounts are required for interactive use.

The OAuth flow:

AI client initiates a connection to the MCP server URL
MCP server redirects to the Nasuni Portal OAuth authorization endpoint
You sign in with your Nasuni Portal credentials
Portal issues an OAuth token scoped to your account
AI client uses this token for all subsequent tool calls
Token refreshes automatically — no manual re-authentication unless your session expires

Role-Based Access Control (RBAC)

Your MCP server access inherits the same roles and permissions as your Nasuni Portal account. This means:

You can only access data you are authorized to see in the Portal
Tool results are filtered based on your account's RBAC policies
Administrative tools (e.g., role and permission queries) respect your permission level
No elevation of privilege is possible through the MCP server

Data Security

Transport encryption: All communication between AI clients and the MCP server uses TLS (HTTPS)
No data storage: The MCP server does not store your infrastructure data — it acts as a pass-through to the Portal API. Data is retrieved on demand and returned to the AI client
Regional endpoints: MCP server endpoints are deployed per region, ensuring your data stays within the appropriate geographic boundary
Session isolation: Each authenticated session is isolated. Tool calls from one user cannot access another user's data

Token Lifecycle

Property	Detail
Authentication method	OAuth 2.0 Authorization Code flow
Client registration	Dynamic Client Registration (DCR)
Token refresh	Automatic; transparent to the user
Session expiry	Governed by Portal session policies
Revocation	Disconnecting the connector in your AI client revokes the token

Tool Reference

Appliance

Tools for discovering and inspecting appliance inventory and configuration.

Tool	Description
`list_appliances`	List all Edge appliances, including build metadata and service enablement flags. Optionally filter by File IQ or S3 Edge availability.
`get_appliance_details`	Get comprehensive details for a single appliance — volumes, machine specs (CPU, RAM, disks), build version, IP address, and location.
`get_appliance_s3_status`	Get S3 Edge service configuration for a single appliance, indicating whether S3 Edge is enabled.
`get_appliance_fiq_service`	Get File IQ (event forwarder) service configuration for a single appliance, including enablement status and connected File IQ appliance.

Volume

Discover and inspect volume details.

Tool	Description
`list_volumes`	Return all volumes available to the caller.
`get_volume`	Return details for a single volume by volume ID.

Nasuni Data Service (NDS) And Snapshot Retention as a Service (SRaaS)

Tools for querying Nasuni Data Services (NDS) configuration.

Tool	Description
`get_nds_config`	Return the Nasuni Data Services (NDS) configuration for a UaaS stack.
`get_nds_stack`	Return detailed configuration for a single UaaS stack (log level, retention, remote support, volume refresh interval).
`list_nds_stacks`	Return all UaaS stacks for the account.
`list_nds_access_keys`	Return blob storage access keys for a UaaS stack.
`get_sraas_status`	Return SRaaS (Storage Retention as a Service) status and run metrics.

Activity logs

Search and retrieve Portal activity logs for audit and compliance purposes.

Tool	Description
`list_activity_logs`	Search Portal activity logs with filters for date range, action type, target type, and free-text search. Returns actor, action, target, timestamp, source IP, and change details.

Roles & Permissions

Query Portal roles, permissions, and your own access level.

Tool	Description
`get_my_roles`	Return the roles and permissions assigned to your account.
`get_role_permissions`	Return the full permission set for a single Portal role.
`list_permissions`	Return every permission available in the Portal system.
`list_roles`	Return all roles defined for the account.

Ops IQ Tools

Appliance Performance

Ops IQ telemetry tools for monitoring Edge appliance resource utilization, cache efficiency, and network performance.

Tool	Description
`get_appliance_memory_utilization`	Return memory telemetry for an appliance over a given time period. Returns detailed breakdown (v10.0.4+) or legacy counters (older builds).
`get_appliance_cpu_usage`	CPU utilization percentage across the appliance.
`get_appliance_cpu_load_average`	1/5/15 minute load averages compared against appliance core count.
`get_appliance_cache_utilization`	Percent of cache capacity consumed over time.
`get_appliance_cache_disk_io`	Read/write I/O metrics for cache disks.
`get_appliance_cache_disk_latency`	Read/write latency for cache disks (in ms).
`get_appliance_cache_hits_misses`	Cache hit/miss distribution — useful for sizing cache tiers.
`get_appliance_network_throughput`	Inbound/outbound throughput across NICs.
`get_appliance_smb_connections`	Concurrent SMB client connection count.
`check_appliance_performance`	Quick appliance performance health check (current state).
`get_appliance_analysis_guide`	Narrative guide for quick, standard, and comprehensive appliance analysis workflows.

Volume-Data Protection and Propagation

Ops IQ tools for monitoring snapshot protection, propagation timing, and cross-appliance sync status.

Tool	Description
`check_appliance_sync_status`	Check a specific appliance's sync status against its volumes. Returns version lag, health status, and source identification.
`list_appliance_sync_statuses`	Return both snapshot details and propagation status for all appliances on a volume.
`get_volume_protection_metrics`	Return snapshot-details telemetry for a volume.
`get_volume_propagation_metrics`	Return snapshot-propagation telemetry for a volume.
`get_volume_latest_version`	Return the latest snapshot version for a volume.
`get_volume_protection_timing`	Return the complete protection cycle timing for a volume.
`get_volume_propagation_timing`	Calculate total file availability timing from source to destination.
`compare_volume_protection_metrics`	Compare data protection metrics (avg, min, max) across multiple volumes. Ranks volumes from best to worst protection.

Volume -Data protection and Propagation Health

High-level health classification and reporting across volumes and the entire fleet.

Tool	Description
`get_fleet_volume_health_summary`	Analyze all volumes and classify each as Healthy, Warning, or Critical.
`get_volume_health_report`	Generate a comprehensive Volume Health Report covering protection and propagation.
`get_volume_data_protection_report`	Generate a formatted Data Protection Report for a volume.

Re-Sizing

Cost analysis and rightsizing recommendations for individual appliances or the full fleet.

Tool	Description
`analyze_appliance_tco`	Analyze the Total Cost of Ownership (TCO) and generate rightsizing recommendations for a single appliance.
`get_fleet_tco_summary`	Fleet-wide TCO analysis and savings recommendations across all appliances.

Performance Diagnostics

AI-driven diagnostic workflows and guided analysis for appliance health and cache performance.

Tool	Description
`guide_appliance_health_quick`	Quick 3-step guidance for appliance health checks.
`analyze_appliance_health`	Comprehensive AI Ops health check: compares current appliance performance against 2-week telemetry baselines to identify anomalies and forecast potential resource constraints.
`analyze_cache_disk_performance`	Analyze cache disk trends and anomalies across SMB connections, cache utilization, hit/miss ratios, IOPS, I/O time, and latency.
`guide_appliance_health_standard`	Standard 6-step guidance for appliance health reports.
`guide_appliance_health_comprehensive`	Comprehensive multi-step guidance for deep appliance analysis.

Troubleshooting

Connection fails during OAuth

Verify your MCP server URL is correct (copy it fresh from Portal → Account Settings → Integrations → Portal MCP Server)
Ensure your Portal account has MCP Server access (licensed feature)
Check that Development/Developer Mode is enabled in your AI client
Try in a private/incognito browser window to rule out cached credentials

Tools are not appearing after the connection

In ChatGPT, scroll down on the connector detail page — tools are listed below the connection info
In Claude, tools appear automatically once the connector is active. Start a new conversation if tools don't appear
In Copilot Studio, verify the connection is active under Settings → Connection Settings

"Unauthorized" or "Forbidden" errors on tool calls

Your Portal session may have expired. Disconnect and reconnect the connector to re-authenticate
Verify your Portal account has the necessary permissions for the data you're requesting. RBAC rules apply to MCP tool calls the same way they apply to the Portal UI

AI client returns unexpected or incomplete results

Some tools require specific parameters (e.g., appliance serial number, volume GUID). Check the tool description for required inputs
For large environments, results may be paginated or summarized. Ask the AI to provide more details if needed
If a tool returns an error, the AI client will typically surface the error message. Retry or rephrase your question