Quick Zap blueprint: trigger on new episode (RSS/Drive/Sheet) → use ChatGPT to assemble the speaking script with intro/outro placeholders → call ElevenLabs (Webhooks or app) to synthesize in short segments → run a serverless ffmpeg step (Lambda/Cloudflare) to concatenate segments + add intro/outro → push final MP3 to hosting. For rate limits: chunk long text, queue with Zapier Storage/Delay, implement exponential backoff on 429s. For cloning: supply clean, varied recordings, consistent mic, consent, and validate on short samples first.