Generate TTS voice lines and multi-speaker dialogue for game NPCs via ElevenLabs.
npx skills add https://github.com/SummerEngine/summer --skill voice-line-summerengine
Please help me install this Agent Skill. Skill: voice-line Source: https://github.com/SummerEngine/summer/tree/main/library/skills/voice-line Command: npx skills add https://github.com/SummerEngine/summer --skill voice-line-summerengine
.summer/memory/casting/voices.md. - Delivery tuning: Maps delivery styles (calm narration, combat bark, robotic, scared) to ElevenLabs voiceSettings parameters like stability, similarityBoost, style, and speed. - Multi-speaker dialogue: Uses the text_to_dialogue capability with a flat inputs array of text and voiceId pairs to produce a single MP3 with dialogue-aware pacing. - Engine wiring: Imports generated audio into the project and wires it to an AudioStreamPlayer on the Voice bus with music ducking. - Use Case: Give a gruff blacksmith NPC a consistent voice across all his lines, generate three bark variants, and wire them into the scene so music ducks when he speaks. ## Quick Start Generate a TTS voice line for my NPC guard saying "Halt, who goes there" with a gruff middle-aged male voice and wire it into the scene on the Voice bus.High-intent search queries and answers about installing and using this skill.
Call summer_generate_audio with capability text_to_speech, the line text, a voiceId, and voiceSettings for stability, style, and speed. Then import the returned MP3 with summer_import_from_url and wire it to an AudioStreamPlayer node.
Use the text_to_dialogue capability with a flat inputs array of { text, voiceId } objects in turn order, repeating the voice id on each line. The result is a single MP3 with dialogue-aware pacing, defaulting to the eleven_v3 model.
There is no voice-listing tool, so browse and audition voices at elevenlabs.io/app/voice-library and paste the voice id. Omitting voiceId falls back to a default voice, which is fine for drafts but not for a shipped cast.
For a combat shout use stability around 0.3, style 0.6, and speed 1.1; for a calm idle bark use stability 0.6 and style 0.2. Lower stability gives more expressive, variable delivery while higher values sound monotone.
Do not use TTS for non-verbal vocal sounds like grunts or screams; use a sound-effect generator instead. Also skip TTS when you have pre-recorded lines from a real actor, which you can import directly as MP3 or WAV.
A wrong or guessed voiceId causes a hard provider error rather than a different-sounding voice. Always source the id from the user or the project's locked cast bible, and never invent voice ids.
shalevaminGenerates ElevenLabs text-to-speech audio with local playback via the sag CLI.
DecioLuvierImplements Godot audio mixing, spatial playback, voice pooling, and interactive music systems.
bivexExtract audio entities from narrative text for sound design analysis.
taiyousan15Convert Japanese text to speech using multiple TTS engines and output formats.