Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
// SPDX-License-Identifier: AGPL-3.0-only
// Copyright 2026-present the Unsloth AI Inc. team. All rights reserved. See /studio/LICENSE.AGPL-3.0
import assert from "node:assert/strict";
import test from "node:test";
import { readSrc } from "./helpers/kit.ts";
const {
buildAudioRunBody,
buildVoiceCreateBody,
sourceFileUrl,
sourceRefOf,
transcribeUrl,
} = await import("../src/features/audio/audio-run-request.ts");
const api = readSrc("features/audio/api.ts");
const PATHLIKE = /"(path|voice_ref|file|url|audio_path)"\s*:/;
test("a clone run sends ids only, with the contract's keys", () => {
const body = buildAudioRunBody({
workflow: "clone",
text: "Hello",
language: "English",
instructions: " ",
inputs: {
reference: { input_id: "a".repeat(32) },
reference_text: " Okay, I'm Cemo. ",
emotion: { clip_id: "c1" },
},
options: { x_vector_only_mode: true, bad: Number.NaN },
speed: 1.3,
seed: 7,
});
assert.deepEqual(body, {
workflow: "clone",
text: "Hello",
language: "English",
inputs: {
reference: { input_id: "a".repeat(32) },
reference_text: "Okay, I'm Cemo.",
emotion: { clip_id: "c1" },
},
options: { x_vector_only_mode: true },
speed: 1.3,
seed: 7,
});
assert.doesNotMatch(JSON.stringify(body), PATHLIKE);
});
test("a cleared transcript is sent blank, an omitted one is left out", () => {
const cleared = buildAudioRunBody({
workflow: "clone",
text: "Hi",
inputs: { reference: { voice_id: "v1" }, reference_text: " " },
});
assert.deepEqual(cleared.inputs, {
reference: { voice_id: "v1" },
reference_text: "",
});
const omitted = buildAudioRunBody({
workflow: "clone",
text: "Hi",
inputs: { reference: { voice_id: "v1" }, reference_text: null },
});
assert.deepEqual(omitted.inputs, { reference: { voice_id: "v1" } });
});
test("whatever a caller spreads in, no path or bytes reach the body", () => {
const sneaky = {
workflow: "speak" as const,
text: "Hi",
path: "/etc/passwd",
voice_ref: "/abs/ref.wav",
inputs: {
reference: { voice_id: "v1", path: "/x.wav", url: "/y" } as never,
emotion: { input_id: "i", clip_id: "c" } as never,
},
};
const body = buildAudioRunBody(sneaky);
assert.deepEqual(body, {
workflow: "speak",
text: "Hi",
inputs: { reference: { voice_id: "v1" } },
});
assert.doesNotMatch(JSON.stringify(body), PATHLIKE);
});
test("a convert run sends the source, the target and the convert settings, and no text", () => {
const body = buildAudioRunBody({
workflow: "convert",
inputs: {
source: { clip_id: "c1" },
target: { voice_id: "v1" },
source_text: " Hello there. ",
},
convert: { mode: "speech", pitch: 3.4, pitch_auto: false, style: "target" },
options: { route: "v2_vc", length_adjust: 1.1, bad: Number.NaN },
seed: 7,
});
assert.deepEqual(body, {
workflow: "convert",
inputs: {
source: { clip_id: "c1" },
target: { voice_id: "v1" },
source_text: "Hello there.",
},
convert: { mode: "speech", pitch: 3, pitch_auto: false, style: "target" },
options: { route: "v2_vc", length_adjust: 1.1 },
seed: 7,
});
assert.equal("text" in body, false);
});
test("a convert run to a built-in voice has no target, and Auto keeps its extra shift", () => {
const body = buildAudioRunBody({
workflow: "convert",
inputs: { source: { input_id: "i1" }, target: null, source_text: " " },
convert: {
mode: "singing",
pitch: 5,
pitch_auto: true,
voice: " manthos ",
},
});
assert.deepEqual(body, {
workflow: "convert",
inputs: { source: { input_id: "i1" } },
convert: {
mode: "singing",
pitch: 5,
pitch_auto: true,
voice: "manthos",
},
});
const low = buildAudioRunBody({
workflow: "convert",
inputs: { source: { input_id: "i1" } },
convert: { mode: "speech", pitch: -40, pitch_auto: false },
});
assert.equal((low.convert as { pitch: number }).pitch, -24);
});
test("whatever a caller spreads into a convert run, no path or unknown key reaches the body", () => {
const sneaky = {
workflow: "convert" as const,
text: "should not be sent",
voice_ref: "/abs/ref.wav",
inputs: {
source: { input_id: "i1", path: "/x.wav" } as never,
target: { voice_id: "v1", clip_id: "c2" } as never,
reference: { voice_id: "v9" },
target_voice: "/abs/target.wav",
} as never,
convert: {
mode: "humming",
pitch: null,
pitch_auto: false,
route: "v1_svc",
source_audio: "/abs/src.wav",
} as never,
options: { retrieval_blend: 0.5, nested: { path: "/x" } } as never,
};
const body = buildAudioRunBody(sneaky);
assert.deepEqual(body, {
workflow: "convert",
inputs: { source: { input_id: "i1" } },
convert: { mode: "speech", pitch: null, pitch_auto: false },
options: { retrieval_blend: 0.5 },
});
assert.doesNotMatch(JSON.stringify(body), PATHLIKE);
assert.doesNotMatch(JSON.stringify(body), /target_voice|source_audio|"text"/);
});
test("selections become one-id refs and server URLs", () => {
assert.deepEqual(
sourceRefOf({ kind: "input", id: "i1", name: "a", durationS: 1 }),
{ input_id: "i1" },
);
assert.deepEqual(
sourceRefOf({ kind: "clip", id: "c1", name: "a", durationS: 1 }),
{ clip_id: "c1" },
);
assert.deepEqual(
sourceRefOf({ kind: "voice", id: "v1", name: "a", durationS: 1 }),
{ voice_id: "v1" },
);
assert.equal(
sourceFileUrl({ kind: "input", id: "i1", name: "", durationS: null }),
"/api/inference/audio/inputs/i1/file",
);
assert.equal(
sourceFileUrl({ kind: "voice", id: "v1", name: "", durationS: null }),
"/api/inference/audio/voices/v1/file",
);
assert.equal(
transcribeUrl({ input_id: "i1" }),
"/api/inference/audio/inputs/i1/transcribe",
);
assert.match(transcribeUrl({ clip_id: "c 1" }), /\?clip_id=c%201$/);
});
test("a saved voice is made from an upload or a clip, within the limits", () => {
const body = buildVoiceCreateBody({
source: { clip_id: "c1" },
name: ` ${"n".repeat(100)} `,
transcript: "",
language: "English",
});
assert.deepEqual(body, {
source: { clip_id: "c1" },
name: "n".repeat(80),
language: "English",
});
assert.throws(() =>
buildVoiceCreateBody({ source: { voice_id: "v" }, name: "x" }),
);
});
test("api.ts sends run and voice bodies only through the builders", () => {
// Named for a speech model kept beside the chat model, which would otherwise get the request.
assert.match(
api,
/body: JSON\.stringify\(withAudioSlotModel\(buildAudioRunBody\(request\)\)\)/,
);
assert.match(api, /body: JSON\.stringify\(buildVoiceCreateBody\(request\)\)/);
assert.match(api, /"\/api\/inference\/audio\/run"/);
assert.match(api, /xhr\.send\(blob\)/);
assert.match(api, /body: blob,/);
assert.doesNotMatch(api, /voice_ref/);
});
test("an edit run sends the source by id and only the edit part's allowed keys", () => {
const markup = 'A <sub targ="robot">human</sub> voice.';
const body = buildAudioRunBody({
workflow: "edit",
text: "A robot voice.",
inputs: {
source: {
input_id: "in_1",
trim: { start_s: 0, end_s: 4 },
path: "/tmp/x.wav",
} as never,
reference_text: " A human voice. ",
},
edit: {
mode: "words",
markup,
instructions: ["", 3 as never],
speed: Number.NaN,
pitch_steps: 1.5,
path: "/tmp/y.wav",
} as never,
});
assert.deepEqual(body, {
workflow: "edit",
text: "A robot voice.",
inputs: {
source: { input_id: "in_1" },
reference_text: "A human voice.",
},
edit: { mode: "words", markup },
});
assert.doesNotMatch(JSON.stringify(body), PATHLIKE);
const editOf = (edit: object) =>
buildAudioRunBody({ workflow: "edit", text: "x", edit: edit as never })
.edit;
assert.equal(editOf({ mode: "bogus" }), undefined);
const delivery = { mode: "delivery", speed: 1.5, pitch_steps: 3 };
assert.deepEqual(editOf(delivery), delivery);
});