Redactor API
Automate privacy workflows with server-side APIs for loading media, detecting sensitive content, exporting redacted output, and integrating Redactor into your existing systems.
[Start with API basics](../examples/api-basics/)
[View video process reference](../reference/videos/process/)
[Handoff to AI Agent](../reference/agent-toolkit/)
Privacy-first workflows
Detect and redact faces, plates, speech, screens, documents, and other sensitive content.
Server API access
Use Redactor Server to integrate automated video and image redaction into internal tools.
Deployment controls
Configure API behavior with environment variables for validation, queues, logging, and storage.
Sighthound Redactor is an application for the redaction of sensitive information in videos and images. It contains state of the art computer vision algorithms for face, vehicle, and speech detection along with object tracking to greatly reduce the amount of operator interaction required to complete redaction tasks. Redactor offers an API that developers can use to add video redaction to their workflows and applications. The API is only available in Redactor Server, not Redactor Desktop.
## Quick Start
- Contact [sales@sighthound.com](mailto:sales@sighthound.com) to obtain a Redactor API serial number
- Install Redactor Server (Windows or Docker)
- Complete the [initial configuration](https://docs.redactor.com/server/admin/initial-setup/#clientserver-account-setup){target=_blank} (authentication and licensing)
- Add the required environment variable `REDACTOR_API_ENABLE=1` to the host instance, and review the list of optional variables below to see if any others are desired (restart recommended)
## System Requirements
Sighthound Redactor has been optimized to efficiently scale in a multicore environment. For best performance, particularly when working with long videos or those that contain many objects, an 8+ core system and at least 2GB of memory per core are recommended. We may recommend certain upgrades if there are concurrent users or for certain workflow considerations. Please reach out to one of our experts at [support@redactor.com](mailto:support@redactor.com) for guidance.
### Recommended Specs (Multiple Concurrent Users)
__OS:__ Linux/Docker or Windows
__CPU:__ 16+ cores. Intel i7, i9, or Xeon (12th gen or newer) or comparable AMD
__RAM:__ 32GB+
__HD:__ SSD with sufficient capacity to store your videos
__GPU:__ NVIDIA GPU w/ 16GB+ RAM and [Compute Capability >= 5.0](https://developer.nvidia.com/cuda-gpus){:target="_blank"}
For help with installing NVIDIA drivers, see our [NVIDIA Driver Installation Guide](https://docs.redactor.com/guides/nvidia-driver-installation/){:target="_blank"}.
__Note:__ Redactor does not yet support acceleration on NVIDIA’s new Blackwell GPUs (e.g., 5070). Tasks will run in CPU mode instead. Support is in development and will be available soon.
__Browser:__ Recent version of Chrome, Firefox, Edge, or Safari
### Virtual Machine Support
Running Sighthound Redactor in a virtual machine (VM) is supported. To ensure compatibility, VMs **must be configured to use a single CPU socket.** You can assign multiple cores to this single socket to meet performance needs, in line with the general system requirements.
For step-by-step instructions on configuring this in popular virtualization platforms, see our [VM CPU Socket Configuration Guide](https://docs.redactor.com/guides/vm-cpu-socket-configuration/){target=_blank}.
## Environment Variables
- `REDACTOR_API_ENABLE` - Required to enable the Redactor API if set to '1'. *Ensure you're licensed for API use.* (default: '0')
- `REDACTOR_API_LOCALFS` - Allows Redactor API to load/save videos on the local filesystem or network using paths such as `file:///example/linux/path/video.mp4` or `file:///C:/example/windows/path/`, if set to `1`. (default: '0')
- `REDACTOR_API_OP_MAX_RUNNING` - How many API operations to run in parallel. (default: '4')
- `REDACTOR_API_OP_QUEUE_LIMIT` - How many API operations to queue if the system is busy. (default: '1024')
- `REDACTOR_API_OP_QUEUE_SYNC` - Queue synchronous API operations (if set to '1'). (default: '0')
- `REDACTOR_API_SPEC_ENABLE` - Enable spec (OpenAPIv3 documentation serving under /apispec/[json|yaml] (if set to '1'). (default: '0')
- `REDACTOR_API_TOKEN` - If specified, all client API calls will require an `Authorization: Bearer {TOKEN}` header.
- `REDACTOR_API_VALIDATE` - To enable Redactor API request/response validation if set to '1' ('2' for development). (default: '0')
- `REDACTOR_FORCE_POLLING_TRANSPORT` - Browser message transport uses polling to mitigate connectivity issues, if set to `1`. (default: '0')
- `REDACTOR_HASH_TYPES_EXPORT` - hash types (md5, sha1, sha256, sha512) used for export/audit logs purposes, comma separated.
- `REDACTOR_HASH_TYPES_IMPORT` - hash types (md5, sha1, sha256, sha512) used for importing, comma separated. (default: 'sha256')
- `REDACTOR_LOG_LEVEL` - Produces more or less log information (debug|info|warn|error). (default: 'info')
- `REDACTOR_TLS_ACCEPT_UNAUTHORIZED` - Accept invalid TLS certificates when set to `1`. (default: '0')
This list is also shown, in the server version, via the `/env/report?dev` route, so you can also make sure that your current version knows about the assumed variables.
## Supported Media Containers and Codecs
__Containers:__ MP4, TS, AVI, MKV, MOV, and MPG
__Video:__ h264, h265, mpeg-4, VP8, VP9, mpeg1video, mpeg2video, msvideo1
__Audio:__ AAC, MP1, MP2, MP3, PCM
## Supported Image Formats
apng, bmp, gif, jp2, jpeg, jpg, png
---
# Agent Instructions
Use this Markdown page as context for Sighthound Developer Portal questions. For broader navigation, read https://dev.sighthound.com/llms.txt. Answer from Sighthound documentation, cite relevant source URLs, and do not ask users to paste secrets, tokens, license keys, or credentials into chat.