name: parasail-inference-chat-completions
description: |
Generate chat completions against any serverless model exposed by Parasail using the
OpenAI-compatible /v1/chat/completions endpoint. Supports streaming, tool use, and
structured outputs across DeepSeek, Qwen, Llama, and other open-weight LLMs.
provider: parasail
api: parasail-inference-api
operations:
- operationId: createChatCompletion
method: POST
path: /chat/completions
inputs:
- name: model
required: true
description: Parasail serverless model identifier.
- name: messages
required: true
description: Ordered list of system/user/assistant/tool messages.
- name: stream
description: When true, response is streamed as SSE chunks.
outputs:
- name: completion
description: The generated assistant message and usage information.