Implementation:Ollama Ollama OpenAPI Spec

Knowledge Sources	Ollama
Domains	API, Documentation
Last Updated	2025-02-15 00:00 GMT

Overview

OpenAPI 3.1.0 specification formally defining Ollama's HTTP API, providing machine-readable documentation for all endpoints, request/response schemas, and data types.

Description

Defines all API endpoints with full request/response schemas in standard OpenAPI format. The specification covers 12 endpoints: /api/generate (text completion), /api/chat (chat completion with tool calling and thinking support), /api/embed (embedding generation), /api/tags (list local models), /api/ps (list running models), /api/show (model details), /api/create (create model from Modelfile), /api/copy (copy a model), /api/pull (download from registry), /api/push (upload to registry), /api/delete (remove model), and /api/version (server version). Includes comprehensive schemas for ModelOptions (sampling parameters like temperature, top_k, top_p, min_p, stop sequences), message structures with role/content/tool_calls, streaming response formats, and tool definitions. Server defaults to http://localhost:11434 with optional bearer token authentication.

Usage

Used by the Mintlify documentation site for interactive API playgrounds and by code generation tools to produce type-safe client libraries. Also serves as the canonical reference for API consumers.

Code Reference

Source Location

• Repository: Ollama
• File: docs/openapi.yaml
• Lines: 1-1491

Signature

openapi: 3.1.0
info:
  title: Ollama API
  version: 0.1.0
servers:
  - url: http://localhost:11434

paths:
  /api/generate:    # POST - Generate text completion
  /api/chat:        # POST - Generate chat message
  /api/embed:       # POST - Generate embeddings
  /api/tags:        # GET  - List local models
  /api/ps:          # GET  - List running models
  /api/show:        # POST - Show model details
  /api/create:      # POST - Create model from Modelfile
  /api/copy:        # POST - Copy a model
  /api/pull:        # POST - Pull model from registry
  /api/push:        # POST - Push model to registry
  /api/delete:      # DELETE - Delete a model
  /api/version:     # GET  - Get server version

Import

# Reference the spec at docs/openapi.yaml
# No import needed - this is a standalone specification file

I/O Contract

Inputs

Name	Type	Required	Description
model	string	Yes	Model name for all generation endpoints
prompt	string	No	Text prompt for /api/generate
messages	array	No	Chat messages for /api/chat
input	string/array	No	Text(s) for /api/embed
stream	boolean	No	Enable streaming responses (default: true)
options	ModelOptions	No	Sampling parameters (temperature, top_k, top_p, etc.)

Outputs

Name	Type	Description
response	string	Generated text (for /api/generate)
message	object	Assistant message with content and optional tool_calls (for /api/chat)
embeddings	array	Embedding vectors (for /api/embed)
models	array	List of model information objects (for /api/tags)
done	boolean	Whether generation has completed
total_duration	integer	Total processing time in nanoseconds

Usage Examples

# Example: Generate endpoint request
POST /api/generate
{
  "model": "gemma3",
  "prompt": "Why is the sky blue?",
  "stream": false,
  "options": {
    "temperature": 0.8,
    "top_p": 0.9,
    "seed": 42
  }
}

# Example: Chat endpoint with tool calling
POST /api/chat
{
  "model": "qwen3",
  "messages": [
    {"role": "user", "content": "What is the weather in Paris?"}
  ],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_current_weather",
        "parameters": {"type": "object", "properties": {"location": {"type": "string"}}}
      }
    }
  ]
}

Related Pages

• Principle:Ollama_Ollama_APIDesign