Prompt
I’m unhappy with Rekognition because it’s not enough for text comments
Latest observation
Nov 2, 2026
OpenAI APIWeb search: off
If your issue is moderation of text comments, Rekognition is the wrong AWS service — it’s for image/video analysis, not comment text.
Better options for text moderation:
-
Amazon Comprehend
- Good for detecting sentiment, key phrases, language, and can help with custom text classification.
- Use it if you want to build your own moderation rules around toxic or risky content.
-
Amazon Bedrock / LLM-based moderation
- Best if you need more nuanced handling of:
- harassment
- hate speech
- self-harm
- sexual content
- spam
- profanity in context
- You can create a moderation prompt or use a moderation model.
- Best if you need more nuanced handling of:
-
Third-party moderation APIs
- Services like OpenAI Moderation, Perspective API, or other content-safety tools may work better out of the box for comments.
-
Custom moderation pipeline
- Common approach:
- basic keyword/profanity filter
- ML-based toxicity detection
- human review for edge cases
- Common approach:
If you want, I can help you design a comment moderation architecture on AWS that does a better job than Rekognition.