# Inferencing

Pegasus APIs for Inferencing.

## OpenAI-compatible chat completions endpoint

 - [POST /inferencing/openai/v1/chat/completions](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/chat_completions_openai_v1_chat_completions_post.md): Creates a response for the given chat conversation.

## OpenAI-compatible embeddings endpoint

 - [POST /inferencing/openai/v1/embeddings](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/create_embeddings_openai_v1_embeddings_post.md): Creates an embedding vector for the input text.

## OpenAI-compatible models endpoint

 - [GET /inferencing/openai/v1/models](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/list_models_openai_v1_models_get.md): Lists available LLMs, VLMs and embedders for inference.

## Anthropic Messages API endpoint

 - [POST /inferencing/anthropic/v1/messages](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/messages_anthropic_v1_messages_post.md): Create a message using the Anthropic Messages API format.

## Anthropic Count Tokens API endpoint

 - [POST /inferencing/anthropic/v1/messages/count_tokens](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/count_tokens_anthropic_v1_messages_count_tokens_post.md): Count tokens for an Anthropic messages request.

## Rerank endpoint

 - [POST /inferencing/uniphore/v1/rerank](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/ranker_uniphore_v1_rerank_post.md): Order documents based on their relevance to the query.

## Nemoguard Jailbreak Detect endpoint

 - [POST /inferencing/uniphore/v1/nemoguard-jailbreak-detect](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/nemoguard_jailbreak_detect_uniphore_v1_nemoguard_jailbreak_detect_post.md): Detect attempts to jailbreak LLMs.

## Ranker endpoint (deprecated) (deprecated)

 - [POST /inferencing/uniphore/v1/ranker](https://docs.businessai.uniphorecloud.com/api-reference/api/inferencing/ranker_uniphore_v1_ranker_post.md): Order documents based on their relevance to the query. DEPRECATED: Use /rerank instead.

