Our GPT

ScholarAI Copilot

Community

Blog

ScholarAI

API Reference

Manage Your Account

Support

Retrieves the full text of an article by its pdf_url. May use a cached entry or read from other data sources before trying PDF parsing.
:param pdf_url: the url of the pdf
:param chunk: the chunk number to retrieve
:return: the chunk of the full text requested with the page number and total number of pages

Read the Full text of a paper

Introduction

Make your own implementation of ScholarAI to work with ChatGPT

Make A GPT

Understand the feature set of ScholarAI and see how the different utilities connect

Feature Set

Everything related to access and authentication

Access

Understand how the ScholarAI API uses prompting

Prompting

Mimics the input and output to the OpenAI Chat Completion API: https://platform.openai.com/docs/api-reference/chat/create

Use the ScholarAI Agent to generate chat completions

Retrieves relevant abstracts and paper metadata by a search. Generates an answer using LLMs if generative_mode is set to true.
For an API meant for faster and more replete data retrieval, use /api/fast_paper_search

Search Papers, for Agents

Retrieves relevant abstracts and paper metadata by a search, with minimal fluff and processing to get searches quickly. Generates an answer using LLMs if generative_mode is set to true.
For an API meant for direct use with agents, use /api/abstracts

Fast Search Papers, for Your Code

Uses embedding model to find section of PDF most relevant for answering a question
:param pdf_url: the url
:param question: the question
:return: the chunk most relevant to answering that question and its source

Ask a question to a PDF

Saves a citation to the user's citation manager

Save a Citation to Zotero

Accept a PDF url or multipart form-data containing a file, and add it to the user's project in the database. If no project is specified, add it to a project named "GPT"

Add papers to project

Creates a project using query params

Accepts a project_name, and asks a question to every paper within. If no project is specified, uses a project named "GPT"

API Documentation

Endpoints

Read the Full text of a paper

Query Parameters

Response