# Features

Accepts JSON payloads via POST request
Returns an estimated ‘uplift’ (how much the measurement has increased) between two versions of something, along with a p-value indicating whether the uplift estimate is to be trusted
Performs much better than traditional AB testing algorithms when multiple experiments are combined
API key authentication
Supports high throughput for enterprise use
Designed for rapid integration with minimal configuration

# Table of Contents

Overview
Getting Started
Request Format
Response Format
Error Handling
Rate Limits
Examples
Support

\--- # Overview The ABSynthesis API provides a simple way to calculate the result of an AB experiment, or set of AB experiments, from structured JSON data. It's designed for back end developers and data scientists to quickly get results from online experiments, and correctly combine those results to maximise the power of experiment programs and report overall uplift from all experiments. The ABSynthesis API processes JSON data submitted in a POST request body. Upon receiving the data, the service analyzes the data to determine whether one version of something (e.g. an application user interface element, a backend website process, a farm planting schedule) performs better than another. # Getting Started ## Base URL ``` https://ABSynthesis.p.rapidapi.com ``` ## SDK Installation Python and R wrappers are available to assist users in implementation To install in Python, use pip in a virtual environment: ``` bash pip install ScientificMicroservices ``` The R installation is hosted on GitHub, so requires the `devtools` package to install quickly: ``` R devtools::install_github("jegar-p/R-ScientificMicroservices") ``` # Request Format ## Endpoint POST /ABSynthesis ## Headers The API is configured to work only with **POST** requests and allows only the headers listed below: | Header | Required | Type | Description | | :---- | :---- | :---- | :---- | | X\_RAPIDAPI\_KEY | Yes | string | API key | | X-RAPIDAPI-HOST | No | string | ABSynthesis.p.rapidapi.com | | Content-Type | Yes | string | Must be application/json | If you make non-POST requests or add headers that are not in the list, you will receive an error from the API. ## Authentication This API uses **API Key** authentication. Include your API key in the request header: ``` X_RAPIDAPI_KEY: YOUR_API_KEY ``` ## Body The request must be an HTTP POST request with a JSON body. The JSON data should be structured as an array of objects, where each object represents an experiment (representing a table of four columns with one row per experiment). Each object should include four named key-value pairs: | Element | Description | | :---- | :---- | | trials\_base | The number of trials in the base version of the experiment. e.g. the number of website visitors in the base version. | | successes\_base | The number of successes in the base version of the experiment. e.g. the number of clickthroughs on a website. | | trials\_variant | The number of trials in the variant (new) version of the experiment. e.g. the number of website visitors in the variant version. | | successes\_variant | The number of successes in the variant (new) version of the experiment. e.g. the number of clickthroughs on a website. | Each object must include at least these four named elements. ### Example Request Body The example data represents a set of online experiments. ```javascript // Four online experiments [ {"successes_base":10,"trials_base":50,"successes_variant":15,"trials_variant":70}, {"successes_base":15,"trials_base":60,"successes_variant":25,"trials_variant":80}, {"successes_base":30,"trials_base":80,"successes_variant":65,"trials_variant":90}, {"successes_base":100,"trials_base":500,"successes_variant":200,"trials_variant":500} ] ``` # Response Format The API returns a JSON response that includes an uplift and a p value. Example Response Body: ```javascript {["uplift": .34, "p_value": 0.04} ``` ## Response Fields ### uplift The uplift value indicates how many more successes the variant would see compared to the base as a percentage (e.g. if a variant saw double the number of successes, uplift \=1). ### p-value The p-value represents how much the uplift calculation can be trusted (this is not statistically accurate, but close enough). In this case, a small number is better \- generally people will trust any p value lower than 0.05. If you are willing to be more ‘risky’ in your decision, you may want to set a more lenient threshold, such as 0.1 . # Error Handling The API uses standard HTTP status codes. Error responses contain a machine-readable code and human-readable message. | Code | Status | Message | Notes | | :---- | :---- | :---- | :---- | | 400 | invalid\_payload | "Invalid or missing 'data' field" | Check the URL of the api in your request, and that the json data submitted is correctly formed. | | 401 | unauthorized | "Invalid API key" | Check your key is sourced correctly, and you are subscribed to the API. | | 500 | internal\_error | “Internal Server Error” | Check content type is ‘application/json’, and the data is a correctly formed json | # Rate Limits | Plan | Requests per Minute | Requests per Month | | ----- | ----- | ----- | | Free | 60,000 | 600 (20/day) | | Pro | | 10,000 | | Ultra | \- | \- | Exceeded limits will return HTTP 429 Too Many Requests. # Examples Using curl to send a POST request: ``` curl --request POST \ --url https://ABSynthesis.p.rapidapi.com/ABSynthesis \ --header 'Content-Type: application/json' \ --header 'x-rapidapi-host: ABSynthesis.p.rapidapi.com' \ --header 'x-rapidapi-key:'$rapidapi_key \ --data '[ {"successes_base":10,"trials_base":50,"successes_variant":15,"trials_variant":70}, {"successes_base":15,"trials_base":60,"successes_variant":25,"trials_variant":80}, {"successes_base":30,"trials_base":80,"successes_variant":65,"trials_variant":90}, {"successes_base":100,"trials_base":500,"successes_variant":200,"trials_variant":500} ]' ``` # Using R: ``` library(httr) library(jsonlite) url = "https://ABSynthesis.p.rapidapi.com/ABSynthesis" payload_df = data.frame( "successes_base" = c(10, 15, 30, 50), "trials_base" = c(50, 60, 80, 90), "successes_variant" = c(15, 25, 65, 90), "trials_variant" = c(70, 80, 90, 150)) payload = toJSON(payload_df) response = jsonlite::fromJSON( rawToChar(httr::POST( url, body = payload, httr::add_headers( 'x-rapidapi-key' =YOUR-API-KEY, 'x-rapidapi-host' = 'absynthesis.p.rapidapi.com', 'content_type' ="application/json"), encode = 'json')$content)) print(response) ``` # Notes for Data Scientists * Only count data (i.e. the count of relevant events such as successes or trials) can be used with ABSynthesis, as this is the most common scenario. Please contact us if you’re interested in other types of data. * The summary provided is intended to be as basic as possible. If you would like more information about your experiments, please contact us for a more advanced version. # Notes for Backend Developers * Ensure your server can handle POST requests with JSON bodies. * Implement proper error logging and monitoring to catch and resolve any server-side issues. * Consider adding authentication and rate limiting to secure and manage the API. # Support For issues, questions, or support requests: **Email**: support@scientificmicroservices.com **Docs**: https://scientificmicroservices.com **Slack**: \#api-support on our community Slack **License:** © 2025 Scientific Microservices ACN 694 721 355. All rights reserved. Use of this API is governed by the Terms of Service. # Changelog v1.0.0 (15/11/24) \- Initial public release v1.2.0 (07/05/25) \- Removed metrics unnecessary for decisions