API Documentation MCP EAP#
Early Access Program (EAP)
The API Documentation MCP is in our Early Access Program (EAP). You can connect and build with it today, but tools, configuration, and behavior may change. Access and usage policies are subject to change, and standard production support commitments do not apply yet. Your feedback helps shape what ships.
The API Documentation MCP gives AI coding agents like Claude Code, Cursor, and Codex searchable access to Workiva’s public API reference and guides. Connect a client that supports the Model Context Protocol (MCP) to find endpoints, scopes, parameters, and examples as you write integration code. You do not need a Workiva account or sign in to read the documentation.
Features#
Your agent can search for relevant endpoints and guides, read full pages with parameters and examples, and explore the available documentation to understand how the APIs fit together.
Running API requests is separate from reading documentation. It requires credentials and permissions you configure separately for your agent.
Your agent can use the documentation to write, review, and update integration code. Executing that code depends on your client’s capabilities and your authorization.
Looking for your Workiva data?
This server returns public documentation only. It cannot access your documents, spreadsheets, or other customer data. Use Workiva MCP to work with your Workiva account through OAuth 2.0 sign in.
Getting Started#
Just the server URL. No Workiva account, API grant, API key, or admin setup is needed to connect to the documentation server.
https://mcp.app.wdesk.com/x/public_api_documentation
Detail |
Value |
|---|---|
Transport |
Streamable HTTP (remote server) |
Authentication |
None |
Tools |
|
Content |
Workiva’s public developer documentation only |
The server is read only. It cannot call Workiva APIs or access customer data.
Why there’s no sign in
The documentation server only returns public documentation and has no access to Workiva customer data. No sign in or credentials are required.
The same public documentation is available in each region. Use the appropriate regional host if your network restricts outbound connections.
Region |
Server URL |
|---|---|
US |
|
EU |
|
APAC |
|
Canada |
|
These instructions name the server workiva-docs and use the first regional URL in the table by default. Choose another regional URL if needed. No authentication headers or tokens are required.
Run this command in your terminal:
claude mcp add --transport http workiva-docs https://mcp.app.wdesk.com/x/public_api_documentation
By default, this adds the server for your current project. Use --scope user to make it available across projects, or --scope project to save a shareable .mcp.json file.
Verify: run claude mcp list, or enter /mcp inside Claude Code, and confirm workiva-docs is connected.
Alternatively, add this to .cursor/mcp.json in your project, or ~/.cursor/mcp.json for all projects:
{
"mcpServers": {
"workiva-docs": {
"url": "https://mcp.app.wdesk.com/x/public_api_documentation"
}
}
}
Verify: open Customize in Cursor and confirm workiva-docs is enabled and its documentation tools are available.
Run this command in your terminal:
codex mcp add workiva-docs --url https://mcp.app.wdesk.com/x/public_api_documentation
Alternatively, add this to ~/.codex/config.toml:
[mcp_servers.workiva-docs]
url = "https://mcp.app.wdesk.com/x/public_api_documentation"
Verify: run codex mcp list to confirm the server is configured. Enter /mcp in the terminal UI to inspect the connection and available tools.
Run this command in your terminal:
gemini mcp add --transport http workiva-docs https://mcp.app.wdesk.com/x/public_api_documentation
Alternatively, add this server to the mcpServers object in ~/.gemini/settings.json. Use httpUrl for Streamable HTTP; url selects SSE in Gemini CLI.
{
"mcpServers": {
"workiva-docs": {
"httpUrl": "https://mcp.app.wdesk.com/x/public_api_documentation"
}
}
}
Verify: run gemini mcp list, or enter /mcp inside Gemini CLI, and confirm the server is connected and its tools are available.
Run this command in your terminal:
code --add-mcp '{"name":"workiva-docs","type":"http","url":"https://mcp.app.wdesk.com/x/public_api_documentation"}'
Alternatively, run MCP: Add Server from the Command Palette, or add this to .vscode/mcp.json. VS Code uses a top-level servers object:
{
"servers": {
"workiva-docs": {
"type": "http",
"url": "https://mcp.app.wdesk.com/x/public_api_documentation"
}
}
}
Verify: run MCP: List Servers from the Command Palette, start workiva-docs, and confirm its documentation tools are available in agent chat.
Use a client that supports remote MCP servers over Streamable HTTP. Add an HTTP server with the regional URL above. Configuration keys vary by client, so follow your client’s documentation. Do not add authentication headers or tokens.
Verify: use your client’s server status or tool list to confirm the connection and find list_pages, read_page, search_docs.
MCP remote server connection guide. Check your client’s MCP documentation for its configuration format and controls.
Once connected, your agent can use these documentation tools:
Tool |
Description |
|---|---|
|
List all available documentation pages. Use read_page to retrieve the full content of a page. |
|
Read the full content of a documentation page by its relative path (e.g. ‘2026-01-01/authentication.md’). Do NOT pass full URIs; only the relative path portion. |
|
Search the API documentation by keyword. Returns matching pages with context snippets. Use read_page to retrieve the full content of a result. |
Using it with your API credentials#
The documentation MCP provides knowledge; API access is separate and optional. Reading the docs and drafting code do not require Workiva API credentials. Running requests against a workspace is a separate action that requires appropriate API credentials, permissions, and a client that can execute code. You control that access and authorize execution; the documentation server cannot call Workiva APIs or access customer data.
Try these prompts#
Start with a documentation question. These prompts do not require Workiva API credentials:
“Check the Workiva docs: how do I obtain an OAuth 2.0 access token?”
“Find the endpoint for writing values to a spreadsheet range and show an example request.”
“What changed in the 2026-01-01 API version?”
You can also ask your agent to apply the documentation to integration code:
“Write a script that lists every spreadsheet in my workspace and exports the sheet names to a CSV.”
Your agent can find the endpoints and pagination rules in the docs and draft the script. Review it before authorizing execution with your separately configured API access.
“Upgrade this script from the 2022-01-01 API to 2026-01-01.”
Your agent can compare the versions in the docs and propose updates to each API call.
“This request returns a 400. Check it against the docs and fix it.”
Your agent can read the reference page and identify changes to the request.
“How do I add a user to a workspace with a specific role?”
Your agent can find the relevant reference pages and explain the steps, with links to its sources.
Your agent can use the documentation tools to search for relevant pages, read their content, and explore available documentation. If it answers without using the tools, ask it to check the Workiva docs and cite the pages it reads.
Better together#
Use llms.txt as an index of the documentation and install Skills to give your agent reusable instructions for common Workiva tasks. The documentation MCP adds searchable access to the docs alongside those resources. These resources are optional; the documentation MCP works on its own.
What it can and can’t do#
Can |
Can’t |
|---|---|
Search and read Workiva’s public API reference and guides |
Call Workiva APIs, execute code, or act on your behalf |
Help your agent compare documentation across API versions |
Access customer data, documents, or workspaces |
Connect compatible clients over Streamable HTTP |
Provide API credentials or bypass Workiva API permissions |
Troubleshooting#
If you see |
Try this |
|---|---|
Server fails to connect |
Check the complete URL and outbound HTTPS access to the selected regional host. For Gemini CLI, use |
Client asks for sign in or a token |
Check that the URL ends in |
Tools are missing after a configuration edit |
Reload the client’s MCP servers or restart the client, then check its server status and enabled tools. |
Agent answers without checking the documentation |
Ask it to search the Workiva docs and read the matching pages, and confirm the documentation tools are enabled in your client. |