Implementation:Ollama Ollama ToChatCompletion

Knowledge Sources	Ollama
Domains	API_Design, Data_Transformation
Last Updated	2026-02-14 00:00 GMT

Overview

Concrete tool for translating Ollama responses to OpenAI chat completion format provided by the openai package.

Description

ToChatCompletion converts a final Ollama api.ChatResponse into an OpenAI ChatCompletion with choices, usage, and finish reason.

ToChunk converts a streaming Ollama response into a ChatCompletionChunk for SSE delivery.

ToCompletion handles the legacy completion format for /v1/completions.

ChatWriter is a custom http.ResponseWriter that intercepts writes from the native ChatHandler, deserializes the Ollama response, translates it, and writes the OpenAI format to the actual HTTP response.

Usage

Invoked automatically by the ChatMiddleware's custom response writer.

Code Reference

Source Location

• Repository: ollama
• File: openai/openai.go (ToChatCompletion, ToChunk, ToCompletion), middleware/openai.go (ChatWriter.Write)
• Lines: openai.go:L262-294 (ToChatCompletion), openai.go:L295-326 (ToChunk), openai.go:L336-357 (ToCompletion), openai.go:L126-134 (ChatWriter.Write)

Signature

func ToChatCompletion(id string, r api.ChatResponse) ChatCompletion
func ToChunk(id string, r api.ChatResponse, toolCallSent bool) ChatCompletionChunk
func ToCompletion(id string, r api.GenerateResponse) Completion

func (w *ChatWriter) Write(data []byte) (int, error)

Import

import "github.com/ollama/ollama/openai"
import "github.com/ollama/ollama/middleware"

I/O Contract

Inputs (ToChatCompletion)

Name	Type	Required	Description
id	string	Yes	Unique completion ID (e.g., "chatcmpl-xxx")
r	api.ChatResponse	Yes	Ollama chat response with message, metrics, done flag

Outputs (ToChatCompletion)

Name	Type	Description
ChatCompletion	ChatCompletion	OpenAI format with ID, Object, Created, Model, Choices, Usage

Usage Examples

OpenAI-Format Response

{
  "id": "chatcmpl-123",
  "object": "chat.completion",
  "created": 1677652288,
  "model": "llama3",
  "choices": [{
    "index": 0,
    "message": {
      "role": "assistant",
      "content": "Hello! How can I help?"
    },
    "finish_reason": "stop"
  }],
  "usage": {
    "prompt_tokens": 12,
    "completion_tokens": 8,
    "total_tokens": 20
  }
}

Related Pages

Implements Principle

• Principle:Ollama_Ollama_OpenAI_Response_Translation

Requires Environment

• Environment:Ollama_Ollama_Go_Runtime