# How to Test Fastio MCP Tools with MCP Inspector Testing Fastio MCP tools with MCP Inspector lets developers validate and debug their agent integrations before deployment. This guide covers connecting to the Fastio MCP server, testing tool calls in the visual interface, and troubleshooting common issues like authentication failures and workspace access problems. Source: https://fast.io/resources/test-fastio-mcp-tools-inspector/ Last reviewed: 2026-02-24 ## What Is the MCP Inspector? The MCP Inspector is an interactive developer tool that lets you test Model Context Protocol servers without integrating them into a full application first. Rather than building your agent and hoping the MCP tools work correctly, you can inspect every tool's behavior, verify authentication flows, and debug issues in a controlled environment. For developers building AI agents that interact with Fastio, the Inspector serves as a helpful development workflow tool. You can call any of the multiple MCP tools exposed by Fastio, inspect their request and response formats, and identify configuration problems before they cause runtime errors in your agent. The Inspector runs as a local web server with a visual interface. It supports multiple transport methods, including Streamable HTTP and SSE, which are the two transports available on Fastio's MCP server endpoints. Helpful references: [Fastio Workspaces](/product/workspaces/), [Fastio Collaboration](/product/collaboration/), [Fastio AI](/product/ai/), and the [Fastio MCP server integration guide for developers](/resources/fastio-mcp-server-integration-developers/) for the broader connection and authentication walkthrough. ## What to check before scaling Testing Fastio MCP tools with MCP Inspector Launch the MCP Inspector using npx without installing anything locally: ```bash npx -y @modelcontextprotocol/inspector ``` The Fastio MCP server is remote and hosted, so there is no local server process for the Inspector to spawn and nothing of Fastio's to install. You connect the Inspector to a URL instead. In the connection pane, choose the Streamable HTTP transport and enter `https://mcp.fast.io/mcp`, or `https://mcp.fast.io/mcp/key` if you want to send a scoped API key as an `Authorization: Bearer` header. The legacy SSE transport is at `https://mcp.fast.io/sse` for clients that need it. The Inspector launches a local web server and opens its interface in your browser. Once connected it lists the tools the server advertises, which is the fastest way to confirm your credentials are working before you wire anything into an agent. Full endpoint and authentication detail lives in the [Fastio MCP server integration guide](/resources/fastio-mcp-server-integration-developers/). ## Testing Fastio MCP Tools Once connected, the Inspector displays four main tabs: Tools, Resources, Prompts, and Notifications. The Tools tab is where you'll spend most of your time when validating Fastio integration. Browse the available tools in the left sidebar. Fastio exposes multiple tools organized by function, covering authentication, workspace management, file operations, shares, and AI chat. Click any tool to see its schema, required parameters, and descriptions. To test a tool, fill in the parameters in the right panel and click "Run". For example, to verify authentication works, call the `auth` tool with action `status` and observe the response. If you're not authenticated, the response indicates whether a session exists and what scopes are available. For testing file operations, you'll need to authenticate first. Use the `auth` action `signin` with email and password, or create an agent account with `auth` action `signup`. Once authenticated, test tools like `storage` action `list` to verify you can access your workspaces. ## Debugging Common Issues Several issues commonly arise when testing MCP tools for the first time. The Notifications pane at the bottom of the Inspector window shows server logs, which helps diagnose problems quickly. **Authentication failures** typically stem from expired tokens or incorrect credentials. JWT tokens last one hour, so if you authenticate and wait before testing, the token may have expired. Re-authenticate by calling `auth` action `signin` again. The `auth` action `status` tool tells you whether your current session is valid. **Missing workspace access** happens when an agent is invited to a workspace but not to the org. Fastio distinguishes between internal orgs (where you are a direct member) and external orgs (accessible only through workspace membership). To discover all accessible orgs, call both `org` action `list` and `org` action `discover-external`. Missing either call means you won't see all available workspaces. **Tool parameter errors** often relate to ID format requirements. Workspaces, organizations, and shares use multiple-digit numeric profile IDs, but most endpoints also accept custom names as identifiers. The `storage` tool accepts folder names like `my-project` instead of numeric IDs, which is easier during development. ## Mocking Fastio Tool Responses One advantage of using the Inspector is the ability to mock responses for testing agent behavior without making actual API calls. While the Inspector itself doesn't include built-in mocking, you can test edge cases by observing how Fastio handles various inputs. For file streaming responses, test with different file sizes to understand how the server handles them. Files under multiple MB return base64-encoded content directly, while larger files return a fallback with an HTTP pass-through URL. This distinction matters for agents that need to handle both cases. For search tool responses, test queries that return zero results, single results, and multiple results. The `storage` action `search` returns matches ranked by relevance. Understanding the response structure helps you build reliable agents that handle all result cases gracefully. Test error responses deliberately by providing invalid parameters. For example, call `storage` action `details` with a non-existent node ID and observe the error format. Your agent should handle these errors gracefully with appropriate fallback behavior. ## Best Practices for Agent Development Use the Inspector throughout your development cycle rather than only at the beginning. Start by verifying basic connectivity and authentication, then test each tool your agent will use before writing any agent code. This approach catches configuration issues early. Keep the Inspector running while developing. When you modify your agent code, re-run the relevant tool calls in the Inspector to verify expected behavior. This rapid iteration loop is faster than deploying your agent and troubleshooting in production. Document which tools your agent uses and in what sequence. Recording the call order makes it far easier to reproduce a failure later, and it is the fastest way to spot an agent making redundant round trips. Finally, design your agent around the plan it will actually run on. Every organization can activate a 14-day Business Trial with a credit card; it runs with Business-plan capabilities for those 14 days. Afterward, pricing uses usage-based credits with per-plan seat and storage limits. Review current limits on the [pricing page](/pricing/) before testing a production workload. ## Frequently asked questions ### How do I debug an MCP server? Use the MCP Inspector to test tools in isolation. Launch it with npx, connect to your server endpoint, then call tools and observe responses. The Notifications pane shows server logs, which helps identify authentication issues, parameter errors, and connection problems. ### What is the MCP Inspector? The MCP Inspector is an interactive developer tool for testing Model Context Protocol servers. It provides a visual interface to call tools, inspect responses, view resources, and test prompts without building a full client application. It's the recommended starting point for debugging MCP integrations. ### How do I connect the Inspector to Fastio? Run `npx -y @modelcontextprotocol/inspector` to launch the Inspector, then connect it to a URL rather than a command. Fastio's MCP server is remote, so select the Streamable HTTP transport and enter `https://mcp.fast.io/mcp`, or `https://mcp.fast.io/mcp/key` if you are sending a scoped API key as a bearer token. ### Why do my tool calls fail with authentication errors? JWT tokens expire after one hour. Re-authenticate by calling `auth` action `signin` with your credentials. Use `auth` action `status` to check whether your current session is valid. For agents, consider using API keys which don't expire. ### Can I test Fastio tools without an account? Only a few tools work without authentication: `auth` actions like `signin`, `signup`, and `email-check`, plus `download` with action `quickshare-details`. Most tools require signing in or setting an API key first. ## About Fast.io Fast.io provides shared workspaces where people and AI agents work on the same files, with built-in semantic search and citation-backed chat over what they hold. Agents reach it through a remote MCP server at https://mcp.fast.io/mcp, a REST API at https://api.fast.io/current/, and a command line client published on npm as @vividengine/fastio-cli.