The original title is: "Four free neural TTS options for CI pipelines — edge-tts, Kokoro, MeloTTS, Bark"

Short summary

Practical comparison of four free TTS options for GPU-free CI pipelines: edge-tts streams broadcast-quality audio via unofficial Microsoft endpoint; Kokoro and MeloTTS offer local CPU inference with slower rendering for videos over 3–4 minutes; Bark excels in expressiveness but requires GPU. Choose based on quality-vs-locality tradeoff.

• •edge-tts: best quality, streams from unofficial Microsoft endpoint, fastest but operational risk
• •Kokoro/MeloTTS: fully local inference, acceptable quality, suitable for short-form video (under 3–4 min)
• •Bark: most expressive with inline effects, requires GPU, impractical for standard CI runners

Generated with AI, which can make mistakes.