ABMeter
Let your coding agent create flags, read experiment results, and run AI-first A/B workflows from ABMeter without clicking the dashboard for every change.
Overview
ABMeter is a Grow-phase MCP server that lets AI agents manage feature flags and A/B experiments through ABMeter’s hosted streamable HTTP endpoint.
What is this MCP server?
- Hosted MCP at https://mcp.abmeter.ai (streamable HTTP)
- Feature flagging plus A/B testing oriented to AI agent workflows
- Publisher docs at abmeter.ai/docs and connect flow at abmeter.ai/lab/connect-mcp
- Open-source repo github.com/abmeter/abmeter (registry metadata)
- Version 0.1.0—experimentation platform, not a static analytics skill
- Registry version 0.1.0
- Remote URL https://mcp.abmeter.ai (streamable-http)
- GitHub source: github.com/abmeter/abmeter
What problem does it solve?
Founders lose velocity toggling experiments in a web UI while their agent could safely apply flag changes and pull results inline.
Who is it for?
Solo SaaS builders who already use or plan to use ABMeter and want agent-driven flag and experiment operations.
Skip if: Pre-launch products with no traffic, or teams that only need read-only Google Analytics without experimentation APIs.
What do I get? / Deliverables
After MCP connect, your agent can drive ABMeter experimentation workflows from the IDE so shipping and measuring variants stays in one loop.
- Connected ABMeter remote MCP in agent config
- Agent-operable feature flag and experiment tooling
- Tighter loop between code changes and measured variant outcomes
Recommended MCP Servers
Journey fit
Experimentation and feature flags matter after you have users and funnels to optimize—canonical shelf is Grow, not first-line Build scaffolding. AB testing and flag analytics are growth analytics work: measuring variants, rollout safety, and conversion impact.
How it compares
Experimentation and feature-flag MCP, not a general SEO or ads reporting connector.
Common Questions / FAQ
Who is ai.abmeter/abmeter for?
It is for indie builders and small teams running ABMeter who want Claude Code, Cursor, or similar agents to work with flags and A/B tests via MCP.
When should I use ai.abmeter/abmeter?
Use it in Grow when you are iterating on funnels, pricing tests, or feature rollouts and need analytics-phase experimentation from the agent.
How do I add ai.abmeter/abmeter to my agent?
Follow ABMeter’s connect guide at https://abmeter.ai/lab/connect-mcp and add the https://mcp.abmeter.ai remote in your agent’s MCP configuration.