> For the complete documentation index, see [llms.txt](https://openindexai.gitbook.io/openindex.ai-api-documentation-1/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://openindexai.gitbook.io/openindex.ai-api-documentation-1/reference/api-reference.md).

# API Reference

We now offer three main endpoints: collections, agents and chat.

[Collections](/openindex.ai-api-documentation-1/reference/api-reference/collections.md) are groups of documents or URLs that OpenIndex.ai will index and embed.

You can then refer to the collection via a `collectionId` parameter when using the `/chat` endpoint.

[Agents](/openindex.ai-api-documentation-1/reference/api-reference/agents.md) are chat bots with memory about previous conversations. Each agent can be created with a default large language model (LLM) (see [supported models](/openindex.ai-api-documentation-1/reference/api-reference/supported-models.md)), model parameters like temperature, topK, topP, frequencyPenalty, repetitionPenalty and tools (plugins).

The [Chat endpoint](/openindex.ai-api-documentation-1/reference/api-reference/chat.md) can be used to retrieve a user's chats and chat with collections and agents. Our backend will take care of persisting the context of the current conversation.

We also have [Postman documentation available](https://documenter.getpostman.com/view/28695742/2s9YRCXWqW#0f902dd3-aead-4a86-b494-458e39eb06d7).