motionrender 0.2.4 → 0.2.5

package/dist/cli.js

@@ -2951,25 +2951,23 @@ function inferChatTask(messages2, options3) {
   return "compose";
 }
 function createProvider(config = {}) {
-  const id2 = (config.provider || process.env.OVE_AI_PROVIDER || (process.env.FREELLMAPI_BASE_URL ? "freellmapi" : "") || (process.env.NVIDIA_API_KEY ? "nvidia" : process.env.GROQ_API_KEY ? "groq" : process.env.ANTHROPIC_API_KEY ? "anthropic" : "")).toLowerCase();
+  const id2 = (config.provider || process.env.OVE_AI_PROVIDER || (process.env.FREELLMAPI_BASE_URL || process.env.FREELLMAPI_API_KEY ? "freellmapi" : "") || (process.env.NVIDIA_API_KEY ? "nvidia" : process.env.GROQ_API_KEY ? "groq" : process.env.ANTHROPIC_API_KEY ? "anthropic" : "") || "freellmapi").toLowerCase();
   switch (id2) {
+    case "mock":
+      return new MockProvider();
     case "freellmapi":
     case "openai":
-      return attachGroqAudio(new FreeLlmApiProvider(config), config);
+      return new FreeLlmApiProvider(config);
+    case "groq":
+      return new GroqProvider(config);
     case "nvidia":
     case "nim":
       return attachGroqAudio(new NvidiaProvider(config), config);
     case "anthropic":
     case "claude":
       return attachGroqAudio(new AnthropicProvider(config), config);
-    case "groq":
-      return new GroqProvider(config);
-    case "mock":
-      return new MockProvider();
-    case "":
-      throw new Error("No AI provider chosen: pass --provider freellmapi|nvidia|groq|anthropic|mock, or set FREELLMAPI_BASE_URL, NVIDIA_API_KEY, GROQ_API_KEY or ANTHROPIC_API_KEY");
     default:
-      throw new Error(`Unknown AI provider "${id2}"; use freellmapi, nvidia, groq, anthropic or mock`);
+      return new FreeLlmApiProvider(config);
   }
 }
 function describeProvider(provider) {
@@ -3259,6 +3257,65 @@ var init_providers = __esm({
         if (!text) throw new Error("FreeLLMAPI chat returned no text. Check active models and upstream keys in your FreeLLMAPI dashboard.");
         return text;
       }
+      async speak(text, options3 = {}) {
+        const response = await request(`${this.baseUrl}/audio/speech`, {
+          method: "POST",
+          signal: timeoutSignal(options3.signal),
+          headers: jsonHeaders(this.headers()),
+          body: JSON.stringify({
+            model: "auto",
+            input: text,
+            voice: options3.voice ?? "alloy"
+          })
+        }, "FreeLLMAPI speech");
+        return { data: new Uint8Array(await response.arrayBuffer()), extension: "mp3" };
+      }
+      async transcribe(audio, filename, options3 = {}) {
+        const form = new FormData();
+        form.append("file", new Blob([audio]), filename || "audio.mp3");
+        form.append("model", "auto");
+        form.append("response_format", "verbose_json");
+        if (options3.language) form.append("language", options3.language);
+        const response = await request(`${this.baseUrl}/audio/transcriptions`, {
+          method: "POST",
+          signal: timeoutSignal(options3.signal),
+          headers: this.headers(),
+          body: form
+        }, "FreeLLMAPI transcription");
+        const data = await response.json();
+        return {
+          text: String(data?.text ?? ""),
+          words: Array.isArray(data?.words) ? data.words.map((w) => ({
+            text: String(w.word ?? w.text ?? ""),
+            start: Number(w.start) || 0,
+            end: Number(w.end) || 0
+          })) : [],
+          cues: Array.isArray(data?.segments) ? data.segments.map((s) => ({
+            text: String(s.text ?? ""),
+            start: Number(s.start) || 0,
+            end: Number(s.end) || 0
+          })) : []
+        };
+      }
+      async image(prompt, options3 = {}) {
+        const size = options3.size ?? "1024x1024";
+        const response = await request(`${this.baseUrl}/images/generations`, {
+          method: "POST",
+          signal: timeoutSignal(options3.signal),
+          headers: jsonHeaders(this.headers()),
+          body: JSON.stringify({
+            model: "auto",
+            prompt,
+            n: 1,
+            size,
+            response_format: "b64_json"
+          })
+        }, "FreeLLMAPI image");
+        const data = await response.json();
+        const b64 = data?.data?.[0]?.b64_json;
+        if (!b64) throw new Error("FreeLLMAPI did not return image data");
+        return { data: Buffer.from(b64, "base64"), extension: "png" };
+      }
     };
     attachGroqAudio = (provider, config) => {
       if (provider.speak || provider.transcribe || !process.env.GROQ_API_KEY) return provider;
@@ -4521,7 +4578,7 @@ function narrationScene(composition, options3) {
   return next;
 }
 async function narrate(provider, options3) {
-  if (!provider.speak) throw new Error(`The ${provider.id} provider cannot make speech; use --provider groq (or add a speak() adapter)`);
+  if (!provider.speak) throw new Error(`The ${provider.id} provider cannot make speech (or add a speak() adapter)`);
   const brief = options3.brief?.trim();
   let text = options3.text?.trim() ?? "";
   if (!text && brief) {
@@ -4568,7 +4625,7 @@ async function narrate(provider, options3) {
   return { audio: audioFile, captions: captionsFile, durationSeconds, words: transcript.words.length, composition, text };
 }
 async function captionAudio(provider, options3) {
-  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe; use --provider groq (or add a transcribe() adapter)`);
+  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe (or add a transcribe() adapter)`);
   const audioFile = resolve6(options3.audio);
   const data = await readFile8(audioFile);
   options3.onProgress?.("Transcribing\u2026");
@@ -4819,7 +4876,7 @@ function editComposition(options3) {
   return composition;
 }
 async function editVideo(provider, options3) {
-  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe; use --provider groq (or add a transcribe() adapter)`);
+  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe (or add a transcribe() adapter)`);
   const clip2 = resolve7(options3.clip);
   const outDir = resolve7(options3.outDir);
   const width = options3.width ?? 1080, height = options3.height ?? 1920, fps = options3.fps ?? 30;
@@ -4927,7 +4984,7 @@ function shortsCandidatePlan(candidate, cut, maxSilence, keepFillers) {
   return cut ? planCuts(slice, { maxSilence, keepFillers, padding: 0.2 }) : planCuts(slice, { maxSilence: Number.MAX_SAFE_INTEGER, keepFillers: true, padding: 0.2 });
 }
 async function shortsFrom(provider, options3) {
-  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe; use --provider groq (or add a transcribe() adapter)`);
+  if (!provider.transcribe) throw new Error(`The ${provider.id} provider cannot transcribe (or add a transcribe() adapter)`);
   const clip2 = resolve8(options3.clip);
   const outDir = resolve8(options3.outDir);
   const width = options3.width ?? 1080, height = options3.height ?? 1920, fps = options3.fps ?? 30;
@@ -5055,7 +5112,7 @@ function imageSize(size) {
   return sizes[size.toLowerCase()] ?? size;
 }
 async function generateImage(provider, options3) {
-  if (!provider.image) throw new Error(`The ${provider.id} provider cannot generate images; use --provider mock (Groq has no image model; or add an image() adapter)`);
+  if (!provider.image) throw new Error(`The ${provider.id} provider cannot generate images (or add an image() adapter)`);
   const generated = await provider.image(options3.prompt, { size: imageSize(options3.size), signal: options3.signal });
   const file2 = resolve9(/\.(png|jpe?g|webp)$/i.test(options3.out) ? options3.out : `${options3.out}.${generated.extension}`);
   await mkdir12(dirname9(file2), { recursive: true });

package/dist/types/packages/ai/src/providers.d.ts

@@ -130,6 +130,12 @@ export declare class FreeLlmApiProvider implements AiProvider {
     private headers;
     resolveModel(task: ChatTask, explicitModel?: string, signal?: AbortSignal): Promise<string>;
     chat(messages: ChatMessage[], options?: ChatOptions): Promise<string>;
+    speak(text: string, options?: SpeakOptions): Promise<AudioResult>;
+    transcribe(audio: Uint8Array, filename: string, options?: {
+        language?: string;
+        signal?: AbortSignal;
+    }): Promise<Transcript>;
+    image(prompt: string, options?: ImageOptions): Promise<ImageResult>;
 }
 export declare class MockProvider implements AiProvider {
     id: string;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "motionrender",
-  "version": "0.2.4",
+  "version": "0.2.5",
   "description": "JSON-first, frame-deterministic motion graphics video engine: render compositions to MP4, WebM, ProRes, GIF or PNG frames from Node or the command line.",
   "license": "MIT",
   "type": "module",
@@ -77,4 +77,4 @@
       "optional": true
     }
   }
-}
+}