Create an Azure AI Content Safety resource in the Azure Portal and retrieve the endpoint URL and subscription key from the Resource Management section
POST to {endpoint}/contentsafety/text:analyze?api-version=2024-09-15-preview with the Ocp-Apim-Subscription-Key header and a JSON body containing the text field and optional categories array (Hate, SelfHarm, Sexual, Violence)
Parse the categoriesAnalysis array in the response; each category includes a severity score on a 0–6 scale with even values (0, 2, 4, 6) as the primary severity levels
Apply thresholds appropriate to your game's rating and audience — for teen-rated games a severity of 2 or above in violence may warrant a block, while adult titles may set higher thresholds
For real-time chat moderation, call the API synchronously per message; the service is optimized for low-latency synchronous use cases
Known gotchas
Text submissions are limited to 10,000 characters; longer inputs must be split into segments and submitted as separate requests
Azure Content Moderator (the legacy service) is deprecated; use Azure AI Content Safety instead — do not build new integrations on the old Content Moderator endpoints
The API returns severity scores, not binary allow/block decisions; you must implement your own threshold logic appropriate to your game's content policy and age rating
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp