The original title is "I Built a One-Curl Audio Translator with Telnyx AI Inference"

Short summary

A developer built a single-endpoint audio translator using Telnyx AI Inference API that chains STT, LLM translation, and TTS in one Flask app. The app handles file uploads, chunks long transcripts at sentence boundaries, and implements per-stage error handling so partial results survive downstream failures. A clone-and-run Python demo is provided with curl examples for translating Spanish audio to English.

• •Single Flask app chains Telnyx STT, AI translation, and TTS for end-to-end audio dubbing
• •Per-stage error handling preserves partial results when downstream calls fail
• •Includes chunking, temp file cleanup, and auto language detection; production upgrade path outlined

Generated with AI, which can make mistakes.