Skip to main content

ABsmartly MCP Server

The ABsmartly Model Context Protocol (MCP) server lets AI assistants — Claude, Cursor, VS Code, Windsurf, Gemini, ChatGPT and any other MCP-compatible client — operate the ABsmartly experimentation platform directly. The AI plans experiments, creates them, starts them, reads results and updates metrics, all through natural-language prompts.

It exposes 230+ commands across 33 groups: experiments, metrics, goals, teams, applications, segments, webhooks and more.

Why use it​

An MCP server turns ABsmartly into a "tool" your AI assistant can call. Instead of switching between your assistant, the ABsmartly UI and the API docs, you work in one place and the assistant does the rest.

Typical workflows that get faster with MCP:

  • Drafting experiments from a description, including templates with goals, metrics, variants and ownership prefilled
  • Bulk operations — "clone this experiment for the EU app, switch the unit type to device_id, and update the audience"
  • Reading state — "what experiments are running this week and what's their status?"
  • Reviewing setups — "find experiments without owners" or "list metrics pending review"
  • Cross-tool agentic work — combine ABsmartly actions with code edits in Cursor / Claude Code, or with Slack / Jira / Notion via other MCP servers

When to use MCP vs the alternatives​

Use caseBest tool
Conversational, ad-hoc operationsMCP server
Application code that assigns variants and tracks goalsClient SDKs
Programmatic automation, CI/CD, custom integrationsWeb Console API
Visual editing of pages for an A/B testLaunchPad browser extension
Manual UI workThe ABsmartly Web Console

MCP isn't a replacement for the SDKs — application code still assigns variants and tracks goals via the SDK. MCP is for the control plane: configuring, launching and reviewing experiments and metrics.

How it works​

The server exposes just 4 tools that give AI assistants access to the full ABsmartly CLI surface:

ToolPurpose
discover_commandsBrowse command groups or search by keyword
get_command_docsGet detailed docs for a specific command
execute_commandExecute any command with typed parameters
get_auth_statusCheck authentication status

A typical interaction looks like this:

"What experiments are running?"
→ discover_commands(group: "experiments")
→ execute_command(group: "experiments", command: "listExperiments",
params: { state: "running" })

Destructive operations, such as start, stop, archive and delete, always ask you to confirm before they run — through an in-client prompt on the hosted server and local stdio server where the client supports it, or in the chat. Experiment creation from templates returns a preview first.

See the Tools Reference for the full surface.

Authentication at a glance​

MethodBest forHow it works
OAuthInteractive use in Claude Desktop, Cursor, VS Code, ChatGPT and cloud agent platformsBrowser-based login via your ABsmartly SAML/SSO or password
API KeyProgrammatic access, CI/CD, headless, or clients that can't use OAuthPass the key in HTTP headers or via the URL
Federated OIDC tokenYour own backend agent acting for users signed in to your identity provider (e.g. Microsoft Entra ID)Send an access token from your identity provider to your instance's /mcp (enabled on request)

All methods are documented in detail in Authentication. OAuth & Client Registration explains the OAuth flow, the endpoints involved and which clients are accepted. Federated OIDC access tokens explains federated tokens, which also work on the REST API, with a step-by-step guide for Microsoft Entra ID.

Where to connect​

URLTransportsNotes
https://mcp.absmartly.com/mcpStreamable HTTPHosted server. Works with every ABsmartly instance.
https://mcp.absmartly.com/sseSSEHosted server, legacy transport for older clients.
https://your-instance.absmartly.com/mcpStreamable HTTPServed by your own ABsmartly instance.
npx @absmartly/mcpstdioRuns locally with your ABsmartly CLI credentials.

What's next​

  • Quick Start — connect Claude Desktop with OAuth and run your first command in under five minutes
  • Setup — per-client setup for every supported MCP client
  • Authentication — API key formats, OAuth, security model
  • OAuth & Client Registration — the OAuth flow with diagrams, endpoints, and how MCP clients are registered and accepted
  • Usage Examples — natural-language prompts that work today
  • Tools Reference — full list of tools, command groups, resources and prompts
  • Troubleshooting — common errors and how to fix them

Source code​

The MCP server is open source under the MIT license: github.com/absmartly/mcp.