# Genesis

Invoke models via Genesis. First [list available models](models.md), copy the model ID to use from the response, then choose an API based on your task.

## Choose a task

| Task | API |
| --- | --- |
| Generate text from messages | [Chat Completions](text-generation/chat-completions.md) |
| Use separate input and output objects | [Responses](text-generation/responses.md) |
| Use Anthropic request format | [Messages (Anthropic)](text-generation/anthropic-messages.md) |
| Convert text to vector embeddings | [Embeddings](text-embeddings/embeddings.md) |
| Rank candidate content | [Rerank](text-reranking/rerank.md) |

Chat Completions, Responses, and Messages use different input and response schemas; do not mix their parsing logic. Refer to each endpoint's page for exact URLs, credentials, parameters, and response formats.

```{toctree}
:hidden:
:maxdepth: 1

List models <models>
Text generation <text-generation/index>
Text embeddings <text-embeddings/index>
Rerank search results <text-reranking/index>
```