Detectivision - 2026 Pricing, Features, Reviews & Alternatives

Detectivision is a cloud-hosted content moderation API that checks text, images, video, audio, and live streams for harmful content through a single endpoint and a single API key. Harmful content categories covered include sexual content, violence, hate, scams, toxicity, and profanity. Each moderation result returns a flagged verdict alongside per-flag confidence scores ranging from 0 to 1, allowing development and product teams to define their own action thresholds rather than relying on binary pass/fail outputs. The platform targets developers and product teams building user-generated content (UGC) platforms, including dating apps, online marketplaces, online communities, and gaming platforms. Plans range from self-serve tiers (Basic at $29/month, Pro at $99/month) to Enterprise with custom pricing, covering startups through large organizations. Enterprise customers also have access to on-premises deployment on their own hardware.

Detectivision supports five content types under one integration. Text moderation checks up to 10,000 characters across 12 languages for insults, hate, racism, sexism, sexual language, self-harm, profanity, and bad habits. Image moderation accepts JPEG, PNG, WebP, BMP, and GIF files up to 10 MB, detecting nudity, violence, hate symbols, and scams including contact details embedded in images. Video moderation offers detect, capture, and censor modes: capture mode returns the timestamp and bounding box of each finding, while censor mode returns a copy of the video with findings obscured; object tracking across frames is included at no additional charge. Audio moderation processes speech in 12 languages and returns findings with start and end timestamps in capture mode. Live stream moderation accepts stream links over RTSP, RTMP, HLS, HTTP, or SRT protocols, as well as direct camera frames, flagging harmful moments with a corresponding image capture at up to 30 frames per second.

A built-in dashboard gives human moderators visibility into flagged content, with support for custom actions such as Approve or Ban, plus free-text notes. Each moderator action triggers a moderation.action webhook back to the customer's system for downstream workflow integration. Asynchronous webhook delivery is available by adding async:true to any request, halving the credit cost for non-urgent checks. Credits are segmented by content type, so exhausting credits for one type does not affect availability for others. Errors and invalid requests consume no credits. A free trial is available on the Basic plan, including 2 months of access with 10,000 text, 10,000 image, and 300 video credits. Enterprise customers can request custom model training for images, video, and live streams.