Extractor Types

Configuration

Extractor Options

Output Configuration

Processing Flow

Best Practices

Limitations

Supported Input Types

Embedding Configuration

Output Formats

Embedding Models

Homepage

Engineering Blog

Multimodal University

Community

Authorization

Mixpeek

Documentation

API Reference

Integrations

Changelog

Use Mixpeek

Talk to Engineers

Configure and customize multimodal feature extraction for different content types

Feature Extraction

Learn how Mixpeek enables powerful multimodal search applications across different types of media

Introduction

Quickstart

Understanding Mixpeek's architecture and capabilities

How It Works

Understanding the advantages of Mixpeek vs building your own search infrastructure

Benefits

Overview of API limits, restrictions, and tier-specific capabilities

Limits

Learn about the fundamental concepts and components of Mixpeek

Overview

Logical groupings of related content for optimized search and management

Collections

Create isolated environments for organizing and managing your search applications

Namespaces

Core building blocks for content management and search in Mixpeek

Assets

Monitor and manage asynchronous operations in your Mixpeek applications

Tasks

Overview of embedding model options from simple to advanced

Available Models

Customize Mixpeek's embedding models for your specific use case

Fine-Tuning

Create and manage hierarchical classifications for multimodal objects

Taxonomies

Discover, organize, and search multimodal features using automatic and manual clustering

Clusters

Create and manage connections between features, nodes, and clusters in your multimodal content

Relationships

Build powerful multimodal search queries across text, images, and videos

Queries

Narrow down search results with powerful filtering capabilities

Filters

Organize search results into logical clusters based on common fields

Grouping

Adjust search result order using feedback signals and popularity metrics

Reranking

BYO Database

TimescaleDB

MongoDB

Supabase

PostgreSQL

Pinecone

Weaviate

Qdrant

Redis

Healthcheck

This endpoint allows you to search features.

Search Features

Get Feature

**Requirements:**
- Required permissions: write

Full Feature Update

Delete Feature

Retrieves a list of features based on
    the provided filters and sorting criteria. 
    If you provide a sort, then pagination isn't supported.

List Features

Ingest Text

Ingest Video Url

Ingest Image Url

Get all taxonomies for the current namespace

List Taxonomies

Register new taxonomies with their descriptions

**Requirements:**
- Required permissions: write

Create Taxonomy

Get the complete taxonomy that contains the specified node

Get Taxonomy

Delete an existing taxonomy and remove all associated node classifications from features.

    This operation:
    - Deletes the taxonomy and all its nodes
    - Removes any node classifications associated with this taxonomy from all features
    - This action cannot be undone
    

**Requirements:**
- Required permissions: write

Delete Taxonomy

Update an existing taxonomy's metadata

**Requirements:**
- Required permissions: write

Update Taxonomy

Get Taxonomy Node

Update an existing taxonomy node

**Requirements:**
- Required permissions: write

Update Node

Starts an asynchronous task to classify features within collections for a given taxonomy.

Classify Features against Taxonomy

Retrieves a paginated list of classification entries with optional filtering.

List Taxonomy Classifications

Delete Classifications

Get Asset

Full Asset Update

Delete Asset

Partial Asset Update

Get asset details including all related features

Get Asset With Features

List Assets

Search Assets

List Collections

Create Collection

Get a collection using either its name or ID

Get Collection

Update a collection using either its name or ID

Update Collection

Delete a collection using either its name or ID

Delete Collection

List Namespaces

Creates a new namespace with the specified configuration

Create Namespace

Retrieve details of a specific namespace using either its name or ID

Get Namespace

Updates an existing namespace's payload indexes, this is a full update

Update Namespace

Deletes an existing namespace using either its name or ID

Delete Namespace

Returns all available models and their configurations, scoped to the organization

List Available Models

Extract Embeddings

Retrieve a task by its ID.
    
    A task may have an expiration time, after which it will still be returned but marked as expired.
    This allows tracking of historical tasks while indicating their current validity state.

Get Task Information

Kill Task

Retrieve all tasks that are not in a complete state (DONE, FAILED, SKIPPED, or CANCELLED)

Option	Type	Description	Default
`interval_sec`	number	Processing interval for videos	`10`
`confidence_threshold`	number	Detection confidence threshold	`0.5`
`embedding_model`	string	Target index for embeddings	`"default"`

Module	Capabilities	Output
`embed`	Vector embeddings	Embeddings array
`transcribe`	Speech-to-text	Text transcript
`describe`	Scene description	Natural language
`detect`	Object/face detection	Bounding boxes

Format	Structure	Use Case
Dense	`{"vector": [float]}`	Standard vector embeddings
Sparse	`{"indices": [int], "values": [float]}`	Efficient high-dimensional data

Feature Extraction

Extractor Types

Text Processing

Image Analysis

Video Processing

Configuration

Extractor Options

Output Configuration

Processing Flow

Best Practices

Limitations

Embedding Models

Supported Input Types

Text & URLs

Files & Encoded

Embedding Configuration

Output Formats

​Extractor Types

Text Processing

Image Analysis

Video Processing

​Configuration

​Extractor Options

​Output Configuration

​Processing Flow

​Best Practices

​Limitations

​Embedding Models

​Supported Input Types

Text & URLs

Files & Encoded

​Embedding Configuration

​Output Formats

Extractor Types

Configuration

Extractor Options

Output Configuration

Processing Flow

Best Practices

Limitations

Embedding Models

Supported Input Types

Embedding Configuration

Output Formats