Elevenlabs

Generates voiceovers, sound effects, and background music through ElevenLabs' API. The documentation is unusually thorough on the details that actually matter: model tradeoffs (multilingual_v2 for reliability, flash/turbo for SSML pauses, v3 for expressiveness but expect retakes), voice settings tuned by style, and the fact that ellipses don't create pauses reliably. Includes a complete Remotion workflow that generates per-scene audio files with a timing manifest, then syncs them to video sequences. The instant voice cloning setup is straightforward (binary file handles, not paths), and the sound effects cap at 22 seconds while music goes up to 5 minutes. Solid if you're building video automation pipelines or need AI audio that doesn't sound like a robot reading a manual.