> ## Documentation Index
> Fetch the complete documentation index at: https://docs.trychainshift.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Core concepts

> The Platform API resource structure and relationships: brands, prompt sets, prompts, runs, and answers

Every resource in the Platform API is scoped to the organization that the API key belongs to.

![Structure of the collection pipeline under an organization (prompt set, prompt, run, answer)](https://mintcdn.com/chainshift/KXnwIlgcCQxzYS5X/images/en/resources-light.svg?fit=max\&auto=format\&n=KXnwIlgcCQxzYS5X\&q=85\&s=7fdcda63dbd9a02567ecfd2a12b19f0b)

## Prompt set and Prompt

A **Prompt** is a single question to run on AI models (up to 1,000 characters). Every prompt belongs to exactly one **Prompt set**, and a set can hold up to 1,000 prompts. Sets group prompts by purpose, such as "30 competitor comparison questions", and let you run a whole set at once in a run request.

## Runs and answers

A **run** (`runs` in the API) is the unit of a run request. You submit prompts (a list of individual IDs or a set) together with a list of models, get back a `runID`, and collection proceeds asynchronously. Each combination of prompts × models × repetitions becomes one **Answer**.

An Answer contains the answer body (`content`) along with the sources the AI cited (`sources`: URL, title, domain) and the fan-out queries the AI issued internally (`fanouts`). For the full structure, see [Retrieving answers](https://docs.trychainshift.ai/en/guides/collecting-answers.md).

## Next steps

> [card] [Running prompts](https://docs.trychainshift.ai/en/guides/submitting-runs.md)
>
> From registering prompts to polling

> [card] [Retrieving answers](https://docs.trychainshift.ai/en/guides/collecting-answers.md)
>
> Retrieve answers with filters and pagination


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.