From Search on the Web to Search in the IDE
Last year we launched AI-driven semantic search for Meraki API docs on developer.cisco.com—but developers live in their IDEs, not the browser. Without the right context, AI assistants in the IDE fall back to outdated or generic knowledge when it generates code.
The DevNet Content Search MCP server is the next step in that evolution: it takes the same semantic search that powers developer.cisco.com and puts it inside your IDE, where your AI assistant can use it in real time. MCP (Model Context Protocol) is a standard that lets IDE AI assistants call live tools—in this case, searching Cisco API documentation—so you get accurate, current context without switching windows. One click to install, nothing to configure, and your IDE can now ground its code suggestions in official Cisco API specs instead of outdated or generic knowledge.
Today the server supports Meraki and Catalyst Center API documentation; we’re expanding to more Cisco products and DevNet content. Want your product’s APIs included? Create an issue to request it.
One-Click Install, Live API Search
The DevNet Content Search MCP server has one-click install for VS Code or Cursor—your environment is ready, nothing to configure. The server doesn’t replace the browser; it brings the right slice of DevNet into the flow of coding.
Once installed, try a query from your IDE chat panel—for example, “Find Meraki L3 firewall API endpoints for configuring traffic rules” or “Find Catalyst Center APIs for device inventory management”—and see the AI assistant pull live documentation and use it in its response.
What’s Under the Hood
The MCP server exposes three tools that your IDE’s AI assistant can call to ground its answers in DevNet:
- Meraki API Documentation Search — Natural language semantic search over Meraki API documentation. Ask in plain language; the server returns matching endpoints with names, descriptions, API path, method, operation ID, and a link to the official doc.
- Meraki API Operation ID Lookup — When you already know the OpenAPI operation ID (e.g.
createNetworkMerakiAuthUser), this tool returns the precise endpoint details and specification. Useful for validation and for generating code that targets a specific operation. - Catalyst Center API Documentation Search — The same semantic search for Catalyst Center APIs: enterprise network management, device onboarding, assurance, compliance, and related use cases.
Responses are structured so that the AI assistant can reliably extract api_path, api_method, api_operation_id, and documentation_url and cite them in generated code. The documentation returned is the same source of truth that powers developer.cisco.com/docs—keeping outputs tied to the real API surface and making it easier for you to verify what the IDE produces.
Beyond Search: Reference Workflows and Prompts
The MCP server is most useful when it’s part of a larger workflow: fetch the latest API details, then generate or refine code in the same flow. To show what that looks like, we provide reference sample workflows and prompt definitions in the GitHub repo—sample configurations for Meraki and Catalyst Center that demonstrate a documentation-first approach: call the MCP search tools before generating code, and tailor output to “simple script” vs “enterprise-style” patterns. Example prompts cover tasks like fetching network IDs, detecting orphan devices, or scheduling firmware upgrades.
These are reference samples only—starting points for you to adapt in your own environment, not a certified or production-ready offering.
A note on AI-generated code: Output from any AI assistant—whether using this MCP server or not—can be incomplete, incorrect, or insecure. Treat generated code the way you would a draft from a junior colleague: review it, test it, and approve it before running it in any real environment. We take this seriously, and it shapes where we’re heading next.
What’s Next
This launch is about one thing: putting the right API context in front of the IDE so that quality starts from a solid base. Where we go from here has two tracks.
1. Broader coverage. Today the server covers Meraki and Catalyst Center APIs. We’re extending it to more Cisco product APIs and other DevNet content—Learning Labs, code samples, sandbox references—so the IDE can use more of DevNet, not just API specs.
2. From reference samples to developer skills. The workflows and prompts we ship today are reference samples. The next step is to evolve them into developer skills: curated, validated configurations with best practices, security guidelines, and coding standards built in. We’re not there yet, and we’ll be transparent about that progression.
Your feedback will shape both tracks—which APIs to add first, which use cases to prioritize, and how we mature the samples. We invite you to try it, share feedback, and tell us which product APIs and use cases you’d like to see next.
Install (one-click): VS Code · Cursor
Docs & feedback: GitHub repo · Request more APIs
