PicPurify

PicPurify is a cloud API for filtering user-submitted images and video in real time. It fits developers building moderation into marketplaces, social products, dating services, blogs, parental-control tools, or safe browsers. The strongest case is broad image screening backed by a permanent free allowance; teams handling several media types or very high burst traffic should weigh its scope and quotas carefully.

Overview

PicPurify focuses on image moderation rather than being a complete content-safety suite. Its API identifies pornographic and suggestively nude material, gore, money, weapons, drugs, hate symbols, obscene gestures, and QR codes. That category range addresses common user-upload risks, while the absence of text and audio moderation means products needing those channels will require another service.

Face detection adds age-majority and gender classifications. These signals can complement an image-review workflow, but they are most useful when a product has a clear reason to apply them. Video is handled by examining individual frames, with one frame per second as the default. Because billing changes with video size and frame interval, video-heavy services need a separate usage estimate rather than treating it as ordinary image volume.

Key features

Image and video handling

The image endpoint accepts JPG/JPEG, WEBP, PNG, BMP, TIFF, RAW, PDF, ICO, and GIF files. A maximum of 4096×4096 pixels or 16 MB covers typical uploads but may require preprocessing for larger originals. The five-requests-per-second image limit is workable for modest traffic; bursty services may need to request a higher limit, which the maker says is possible for legitimate reasons.

Integration and customization

Examples for cURL, Python, Node.js, and PHP give teams in those ecosystems a direct starting point for connecting a website, app, or service. PicPurify also offers tailor-made models for specific detection requirements, a useful route when the standard categories do not match a specialized marketplace or community. Custom work is handled through the maker rather than a standard self-serve feature.

Pricing

PicPurify uses units: every task performed on one image consumes one unit, so applying several checks to the same image multiplies usage. The free plan costs nothing and includes 2,000 free units with no time limit. There are no extra fees for account setup, onboarding, training, or technical support, making the free tier suitable for evaluation and consistently low-volume projects.

STARTER suits a small production integration, while PRO gives a larger monthly buffer. ULTRA, MEGA, and SUPRA are aimed at higher-volume services and publish declining prices for additional 1,000-unit blocks. STARTER and PRO do not add units by default, so workloads that exceed their quotas need a plan change or a discussion with the maker. Custom pricing is available for specific needs.

Platforms

PicPurify is offered as a cloud API and web service. It is not a self-hosted deployment, so organizations that must keep moderation processing on their own infrastructure should look elsewhere. The documented examples cover cURL, Python, Node.js, and PHP.

Who it's for

This is a practical fit for e-commerce and marketplace listings, social networks, dating sites, blogs, parental-control products, and safe browsers that primarily need image checks. Developers can start without a paid commitment, then choose a unit tier as upload volume becomes predictable. It is a weaker fit for products requiring text or audio moderation in the same platform, or for teams whose traffic regularly exceeds five requests per second without a confirmed increase.

Pros and cons

Pros

• Permanent free allowance: 2,000 units do not expire, supporting ongoing low-volume use rather than a short evaluation window.
• Broad image categories: Coverage includes sexual content, gore, weapons, drugs, hate symbols, gestures, money, and QR codes.
• Developer-friendly entry points: cURL, Python, Node.js, and PHP examples reduce integration friction.
• Custom model option: Tailor-made detection can address needs beyond the standard categories.

Cons

• Unit multiplication: Multiple tasks on one image consume multiple units, making broad checks costlier than upload counts suggest.
• Media scope is limited: There is no text or audio moderation.
• Video budgeting is complex: Frame interval and video size both affect charges.
• Traffic ceiling: Five requests per second may constrain bursts until a higher limit is approved.

Alternatives

Tisane API is a better match for teams wanting self-hosted and desktop operating-system options, though its Community plan limits usage to 50,000 requests and omits technical support, CORS support, direct language-model access, and customization.

Google Cloud Vision API suits organizations preferring pay-as-you-go image features billed per image and feature, with a 1,000-image free allowance in its listed tier.

WebPurify is worth considering when a paid service with live moderation and Android, iOS, API, web, and self-hosted options is preferable.

Amazon Rekognition Content Moderation fits AWS users who can use its 12-month free tier, including 60 free video-analysis minutes per month.

Detoxify is a free, self-hosted open-source library for toxic-comment classification.

Hive Moderation targets paid developers with access to more than 10 Hive models, pay-as-you-go selected models, and custom-model options.

Sightengine offers freemium image moderation with a 1-request-per-second free tier and a paid Starter plan; its free tier excludes video processing and live-stream support.

Llama Guard is a free self-hosted option built around the Llama Guard 3-8B model weights.

For a wider comparison, see AI Content Moderation Software.

Verdict

Choose PicPurify if your product needs broad image and frame-based video moderation through a cloud API, especially when a no-expiry 2,000-unit allowance and familiar language examples matter. Look elsewhere if you need text or audio checks, self-hosting, or a pricing model that does not multiply charges for several tasks on one image.