Skip to content

Redactor Agent Toolkit

The Redactor Agent Toolkit is a distribution package for developers integrating with the Redactor API. It includes Docker Compose deployment configurations, AI agent-optimized documentation, TypeScript type definitions, and the OpenAPI specification.

Obtaining the Toolkit

Contact sales@sighthound.com to obtain the toolkit zip file and an evaluation serial number.

Folder Structure

After extracting the zip, the toolkit contains the following:

redactor-agent-toolkit/
├── AGENTS.md                          # AI agent context (cross-IDE)
├── CLAUDE.md                          # Claude Code configuration
├── .cursorrules                       # Cursor configuration
├── .github/copilot-instructions.md    # GitHub Copilot configuration
├── VERSION.txt                        # Version and build info
├── ai-context/                        # AI-optimized API documentation
│   ├── index.md                       # Navigation index
│   └── ...                            # Markdown reference files
├── reference/                         # API reference materials
│   ├── openapi-spec.yaml              # OpenAPI v3 specification
│   └── sh-types/                      # TypeScript type definitions (.d.ts)
├── deployment-examples/               # Docker deployment
│   ├── docker-compose.yml
│   ├── config/                        # NGINX, SSL, and example HTML files
│   └── volumes/
│       ├── data/                       # Mounted at /data/ in the container (output location)
│       └── redactor/                   # Application state and log files
└── README.md

deployment-examples/

Docker Compose configurations with profiles for different deployment scenarios:

  • API-only — Run Redactor with the API enabled (docker compose up -d)
  • Same-domain proxy — NGINX reverse proxy on the same domain (docker compose --profile with-same-domain-proxy up -d)
  • Cross-domain proxy — NGINX reverse proxy on a separate domain (docker compose --profile with-cross-domain-proxy up -d)

See Embedded Redactor UI for details on the proxy deployment options.

ai-context/

Markdown documentation optimized for AI coding agents and LLMs, so you can ask your agent questions about Redactor as you work. Start at ai-context/index.md for navigation to all API reference, examples, and integration guides. See Using with AI Coding Agents below for setup.

reference/

API reference materials including the OpenAPI specification (openapi-spec.yaml) and TypeScript type definition files (sh-types/). These files are provided for use in your own projects — import the types package or use the OpenAPI spec with your API client generator or documentation tool.

Quick Start

Apple Silicon Not Supported

The Redactor server currently requires an Intel/AMD (x86_64) processor. M-series Macs (Apple Silicon) are not supported at this time.

  1. Extract the toolkit zip to your computer
  2. Open the redactor-agent-toolkit/ folder in your IDE
  3. Run /get-started for an interactive walkthrough that handles everything: starting the containers, configuring Redactor, and making your first API call

Your AI coding agent will automatically discover the Redactor API documentation via AGENTS.md and can start and stop the Docker containers, walk you through setup, answer API questions, and generate integration code.

Manual Setup

  1. Extract the toolkit zip to your computer
  2. cd into the redactor-agent-toolkit/deployment-examples folder

  3. Start the Redactor container:

    docker compose up -d

  4. If this is your first time running this instance of Redactor, set up your initial admin user and add a license at http://localhost:9000/admin

Redactor is now running and accessible at http://localhost:9000.

Volume Mounts

The Docker Compose configuration mounts the following directories from your host into the Redactor container:

./volumes/data/ — Mounted at /data/ inside the container. This is used as an output location for redacted videos and other files generated by the API. Several of the API examples use this path, so you can easily access results from your host machine.

./volumes/redactor/ — Stores all application state and log files. If you encounter any issues, review the log file at ./volumes/redactor/logs/main.log or send it to support@redactor.com for assistance.

./volumes/sio/ — Used for GPU model cache data when running with NVIDIA GPU acceleration.

Using with AI Coding Agents

The ai-context/ folder contains Markdown documentation specifically structured for AI coding agents, giving them full context on the Redactor API so you can ask questions and get help as you build your integration.

Supported agents

  • Claude Code — Reads CLAUDE.md and AGENTS.md. Provides /get-started and /redactor slash commands.
  • OpenAI Codex CLI — Reads AGENTS.md
  • Cursor — Reads .cursorrules
  • GitHub Copilot — Reads .github/copilot-instructions.md
  • Other AI agents — Most AI-powered tools will discover AGENTS.md

Working directly in the toolkit

If you're working from within the redactor-agent-toolkit/ folder, the included AGENTS.md at the root means most AI coding agents will automatically discover the documentation — no extra configuration needed.

Adding the toolkit to your project

If you copy the redactor-agent-toolkit/ folder into your own project's source code, agents will discover AGENTS.md automatically. For other agents, add a reference to the documentation in your project's AI agent configuration file:

Cursor — add to .cursorrules or .cursor/rules/*.mdc:

For Sighthound Redactor API integration, read docs/redactor-agent-toolkit/AGENTS.md

GitHub Copilot — add to .github/copilot-instructions.md:

For Sighthound Redactor API integration, read docs/redactor-agent-toolkit/AGENTS.md

Windsurf — add to .windsurfrules:

For Sighthound Redactor API integration, read docs/redactor-agent-toolkit/AGENTS.md

Adjust the path to match where you placed the toolkit in your project.