comfyui-mcp/dist/services/generate-audio.js

Local-first, agent-native control plane for ComfyUI — MCP server + autonomous sidebar agent that drives your live graph in natural language on ANY LLM: Claude/ChatGPT/Gemini on your subscription (no API key), free local models via Ollama (fully offline),

import { createWorkflow } from "./workflow-composer.js";
import { DefaultsManager } from "./defaults-manager.js";
import { ValidationError } from "../utils/errors.js";
const DEFAULTABLE_KEYS = [
    "seed",
    "steps",
    "cfg",
    "sampler",
    "scheduler",
    "filename_prefix",
    "unet",
    "vae",
    "clip_a",
    "clip_b",
    "lyrics",
    "language",
    "musical_key",
    "shift",
    "guidance_scale",
    "bpm",
    "timesignature",
    "temperature",
    "top_p",
    "top_k",
    "min_p",
    "generate_audio_codes",
    "audio_quality",
    "checkpoint",
    "clip",
    "negative_prompt",
];
export async function generateAudio(args, deps) {
    if (!args.prompt) {
        throw new ValidationError("prompt is required for audio generation");
    }
    if (!args.duration || args.duration <= 0) {
        throw new ValidationError("duration must be a positive number (in seconds)");
    }
    const seed = {};
    for (const key of DEFAULTABLE_KEYS) {
        const v = args[key];
        if (v !== undefined)
            seed[key] = v;
    }
    const resolved = DefaultsManager.apply(seed);
    let checkpoint = resolved.checkpoint;
    let unet = resolved.unet;
    let vae = resolved.vae;
    let clip = resolved.clip;
    let clipA = resolved.clip_a;
    let clipB = resolved.clip_b;
    if (args.model_family === "ace_step_1.5") {
        if (!unet)
            unet = await deps.resolveFirstModel("diffusion_models");
        if (!vae)
            vae = await deps.resolveFirstModel("vae");
        if (!clipA)
            clipA = await deps.resolveFirstModel("text_encoders");
        if (!unet) {
            throw new ValidationError("No UNet model specified or found locally for ACE Step 1.5. " +
                "Pass `unet` or download one via download_model.");
        }
        const workflow = createWorkflow("ace_step_15", {
            unet,
            vae: vae ?? "ace_1.5_vae.safetensors",
            clip_a: clipA ?? "qwen_0.6b_ace15.safetensors",
            clip_b: clipB ?? "qwen_4b_ace15.safetensors",
            prompt: args.prompt,
            lyrics: resolved.lyrics,
            duration: args.duration,
            seed: resolved.seed,
            steps: resolved.steps,
            cfg: resolved.cfg,
            sampler_name: resolved.sampler,
            scheduler: resolved.scheduler,
            shift: resolved.shift,
            language: resolved.language,
            musical_key: resolved.musical_key,
            guidance_scale: resolved.guidance_scale,
            bpm: resolved.bpm,
            timesignature: resolved.timesignature,
            temperature: resolved.temperature,
            top_p: resolved.top_p,
            top_k: resolved.top_k,
            min_p: resolved.min_p,
            generate_audio_codes: resolved.generate_audio_codes,
            audio_quality: resolved.audio_quality,
            filename_prefix: resolved.filename_prefix,
        });
        const { prompt_id, queue_remaining, rejectedOutputs } = await deps.enqueue(workflow);
        return { prompt_id, queue_remaining, rejectedOutputs, model_family: "ace_step_1.5" };
    }
    // Stable Audio 3
    if (!checkpoint) {
        const models = await deps.resolveFirstModel("checkpoints");
        if (models)
            checkpoint = models;
    }
    if (!clip)
        clip = await deps.resolveFirstModel("text_encoders");
    if (!checkpoint) {
        throw new ValidationError("No checkpoint specified or found locally for Stable Audio 3. " +
            "Pass `checkpoint` or download one via download_model.");
    }
    const workflow = createWorkflow("stable_audio_3", {
        checkpoint,
        clip: clip ?? "t5gemma_b_b_ul2.safetensors",
        prompt: args.prompt,
        negative_prompt: resolved.negative_prompt,
        duration: args.duration,
        seed: resolved.seed,
        steps: resolved.steps,
        cfg: resolved.cfg,
        sampler_name: resolved.sampler,
        scheduler: resolved.scheduler,
        filename_prefix: resolved.filename_prefix,
        // #1458 — passed through like the ace_step path. The audio generation tool
        // ACCEPTS audio_quality; this family dropped it on the floor.
        audio_quality: resolved.audio_quality,
    });
    const { prompt_id, queue_remaining, rejectedOutputs } = await deps.enqueue(workflow);
    return { prompt_id, queue_remaining, rejectedOutputs, model_family: "stable_audio_3" };
}
//# sourceMappingURL=generate-audio.js.map