VISIONContent moderation guides
Practical guides to moderating what users post: how to choose an API, what moderation costs, how to handle hate speech, explicit images, video and live streams, and where people still need to decide. Each guide answers one question and shows the requests and answers of Detectivision's API where they help. The newest guides come first, grouped by topic.
Moderate Firebase messages and uploads: a Cloud Function sends each new text or image to a moderation API and keeps it hidden until it is approved. With code.
What OpenAI, Google, AWS, Azure, Sightengine and others give for free, the limits, and the catch, from each vendor’s own pages with the date read.
A reference design for moderating user content: where to check, when to block or hold, how to route doubtful cases to people, and how decisions reach your app.
Apple's Guideline 1.2 asks UGC apps for four things: filtering, reporting, blocking and contact details. How to fix each, and what to tell App Review.
Check Supabase posts and uploads in an Edge Function, keep them hidden with row-level security until approved, and take results by webhook. With code.
Yes, for most of it: a model reads every message and says which kind of abuse it finds and how sure it is. What it does well, where people decide.
Check every image when it is uploaded, then block, blur or hold it when it shows nudity or sexual content. Step by step, with the API call.
Your options for moderating large volumes of user-generated video: sample frames, check asynchronously and send people only what is flagged.
Rekognition, Azure, Google Cloud, Hive, OpenAI, Sightengine, WebPurify and Detectivision side by side: what each checks, languages, and published prices.
Why a SaaS product pays per check for moderation: costs follow usage, nothing to run, failed checks are free. And when another way fits better.
The cheapest way to moderate live streams: check the frames you need, only the flags you act on, and pay for frames checked. With worked examples.