Get Started

Introduction

Introduction

PureFrame is a video intelligence API. Upload your footage and PureFrame indexes every frame, sound, and spoken word — then find any moment with a natural language query or a reference image.

Access these core capabilities through simple REST APIs:

  • Search — find specific moments using text or image queries, across visual content, speech, and audio
  • Organize — group videos into collections and scope searches to what matters
  • Integrate — give any AI agent vision over your video library via MCP or function calling

Jump in

Quickstart

Upload your first video and run your first search in under 5 minutes.

API Reference

Complete documentation for every endpoint.

Agentic Vision

Give Claude, GPT-4o, or any LLM the ability to search your video library.

Collections

Organize videos and scope searches to a specific library.

How PureFrame works

1

Upload

Send a video file to POST /v1/upload. PureFrame returns a job_id immediately — processing happens in the background.

2

Index

PureFrame extracts frames every 5 seconds, embeds them with CLIP, and transcribes speech with Whisper. The job status moves from processing → done.

4

Use the clips

Each result includes thumbnail_base64 — a base64 JPEG that vision-capable models like Claude and GPT-4o can consume directly, no URL fetching needed.

Base URL

https://api.pureframe.ai

All endpoints are versioned under /v1.

Response envelope

Every response shares the same structure:

1{
2  "data": { ... },
3  "meta": { "total": 10, "page": 1, "per_page": 10 },
4  "ok": true,
5  "latency_ms": 42
6}

Errors return a non-2xx status with a machine-readable code:

1{
2  "detail": {
3    "code": "INSUFFICIENT_CREDITS",
4    "message": "Human-readable description"
5  }
6}

See Errors for the full list of error codes.