Published Updated
Introducing SciGround, starting with study design
SciGround connects research design, data, analysis, and claims through mathematical verification. The free public beta starts with registering and checking a planned two-group study in Claude Web.
When AI helps analyze an experiment, you still need to know whether the data match the study design, whether the calculation is right, and what the result supports. We built SciGround to connect those checks, from the original research question to the claim you report.
The first public beta of SciGround MCP is free. It starts where that chain begins: with a planned study. Connect SciGround to Claude Web, sign in with Google, and register a comparison between two independent groups.
Check your design before collecting data
Say you plan to compare a treatment group with a control group. Tell Claude what you want to measure and what counts as one independent sample. Review the details before approving registration. SciGround can flag missing information or problems in the registered settings, so you know what to clarify before collecting data.
Keep your study ID. With the same Google account, you can use it to return to your study, check its current state, and review its diagnostics.
See how the checks connect
SciGround Foundations walks through an illustrative study from design to claim. It shows how the planned comparison guides the data checks, how the calculation is verified, and how a versioned Record lets others inspect the evidence behind the claim.
Public Beta 0.1.0 covers design registration, study state, and design diagnostics. File uploads, statistical analysis, result and Record retrieval, and sharing require access beyond this public beta.
Connect your AI
In Claude Web, add SciGround as a custom connector and sign in with Google. Then ask Claude, “I want to register a study comparing two groups.” Claude’s own plan and workspace requirements apply.