This dataset provides 1,000 synthetic child protection helpline call transcripts paired with concise, structured summaries for training sequence-to-sequence summarization models. It simulates real-world calls reporting various forms of abuse and exploitation in East African contexts while ensuring no real cases or personal data are included