Skip to main content
You’re making a product decision and you need to know what your users think. Semilattice gives you an AI model of your audience that answers questions like they would, grounded in real research data, returned in seconds. From general consumers to software developers to niche segments you define with your own data.

​
How it works

  1. Choose a population model - Select from ready-to-use public population models or create your own
  2. Predict answers - Simulate how the population would answer questions to predict behaviour
  3. Orchestrate predictions - Combine multiple predictions to produce emergent outputs

​
Quick start

Choose your preferred language to get started:

Python

Get started with our Python SDK and start predicting in minutes.

Node.js

Get started with our Node.js SDK and integrate with your JavaScript applications.

​
Core concepts

Populations

Population models represent specific user profiles, such as your customers or new users.

Predictions

Predictions simulate population answers to new questions, generating net new information.

Tests

Tests simulate population answers to known questions, providing accuracy benchmarks.

​
Example use cases

Enable AI agents to predict user responses for automated A/B testing, personalisation, or content optimisation decisions.
Automate user research workflows by simulating responses from specific user profiles at scale.
Build systems that automatically select messaging, features, or experiences based on predicted user profile responses.
Integrate user response predictions into product development pipelines for data-driven feature prioritization.
Create systems that continuously monitor and predict market responses to inform strategic decisions.

API Reference

Complete API documentation with request/response examples for all endpoints