Amplitude experiments inside the product.
Xi runs experiments outside it.
Amplitude Experiment is a product experimentation platform tied to Amplitude analytics: feature flags, in-app variants, statistical engines wired to your product event stream. Xi captures a different layer — the contract behind a marketing experiment, channel-agnostic, no analytics dependency.
Why teams switch from Amplitude Experiment to Xi
Amplitude Experiment is built for product teams already living in Amplitude — same event stream, same dashboards, same analytics paradigm. The fit is excellent for in-app feature experiments. It is a poor fit for marketing experiments because the unit of work is wrong: a flag exposure tied to a product event, not a contract about what you are trying. Teams move to Xi for four reasons.
- No analytics dependency. Xi works with whatever metric source you trust — your CRM, your ad platform, a manual log. No need to model the event in Amplitude first.
- Channels Amplitude does not see. Paid ads, outbound emails, pricing changes — Xi handles them all because it captures the contract, not the exposure.
- Plain language, not flag wiring. A marketer can write a Xi contract in five minutes. Amplitude Experiment requires flag instrumentation and event modeling in the product code.
- The agent runs it. Native MCP server. Claude can commit and decide experiments without touching the SDK.
If you are running product experiments inside Amplitude, keep Amplitude Experiment. If you are running marketing experiments, switch.
Where the two tools actually differ.
When each tool is the right call.
- You are an Amplitude customer running experiments tied to product features.
- You want flag-based variants instrumented to your existing event taxonomy.
- Your experiments live inside the application and depend on product analytics.
- You have an engineering team that owns the flag stack.
- Your experiments are marketing decisions, not flag exposures.
- You do not use Amplitude, or your metrics live elsewhere.
- You want a contract per experiment and a verdict at the end.
- You want the agent to commit experiments via MCP.
Common questions, short answers.
How is Xi different from analytics-tied experimentation?
Different unit of work and different decision rule. Analytics-tied tools measure flag exposures against your product event stream. Xi captures contracts about marketing decisions — channel-agnostic, no analytics dependency.
Can I use Xi alongside Amplitude Experiment?
Yes. Use Amplitude Experiment for in-app product experiments tied to your event taxonomy. Use Xi for marketing experiments and the contract archive.
Does Xi require Amplitude?
No. Xi has no analytics dependency. It works with whatever metric source you already trust.
Do I need to instrument events for Xi?
No. Xi reads metric values from whatever source you choose; there is no event taxonomy to maintain.
Take one idea. Turn it into an experiment.
Free plan, unlimited archive, no card required. See it in Claude / Cursor / Codex in 30 seconds.