@acedatacloud/sdk 2026.1108.0 → 2026.1113.1

package/README.md

@@ -143,6 +143,31 @@ Streaming and `claude.chat.messages.countTokens(...)` also preserve thinking
 configuration. Parameter pass-through does not guarantee model support; API errors
 are returned through the SDK's normal error handling.
 
+### Claude Fable 5.2
+
+Use the exact model ID `claude-fable-5-2`. Native Messages preserves thinking and
+tool content blocks. Omit `temperature`, `top_p` and `top_k`; adaptive thinking is
+always enabled, and the output budget covers thinking plus the final answer.
+For tools, use `tool_choice: { type: 'auto' }`, not `any` or a forced tool.
+Fable 5.1 and Fable 5 remain separate model IDs; upgrade by changing `model` explicitly.
+Count Tokens and Responses are not published for this model. Consult the
+[current model page](https://platform.acedata.cloud/models/claude-fable-5-2?m=chat)
+for account availability and package-dependent prices; model discovery alone does
+not guarantee a successful generation.
+
+```typescript
+const fable = new AceDataCloud({
+  apiToken: 'your-token',
+  baseURL: 'https://api.acedata.cloud',
+});
+const response = await fable.chat.messages.create({
+  model: 'claude-fable-5-2',
+  messages: [{ role: 'user', content: 'Explain this code.' }],
+  maxTokens: 4096,
+  output_config: { effort: 'high' },
+});
+```
+
 ## OpenAI Embeddings
 
 ```typescript
@@ -304,6 +329,54 @@ not trigger reference-video billing. Do not combine `videoList` with
 
 ## Suno music processing
 
+### Replace a section
+
+Use the completed song's `data[].id`, not the task ID. Listen to the source and
+use `await client.suno.timing({ audioId })` to choose a complete phrase; replace
+the example ID and 30–50 second interval with your own. Supply only the local
+new lyrics, preferably the same text in `lyric` and `prompt`, not the whole song.
+
+```typescript
+const localLyrics = 'Evening winds along the quiet lane\nCarry our story home again';
+const task = await client.suno.generate({
+  action: 'replace_section',
+  audioId: 'YOUR_COMPLETED_AUDIO_ID',
+  model: 'chirp-v5-5',
+  lyric: localLyrics,
+  prompt: localLyrics,
+  replaceSectionStart: 30.0,
+  replaceSectionEnd: 50.0,
+  replaceSectionResultMode: 'full_song',
+});
+const result = await task.wait();
+const songs = (result.response as Record<string, unknown>).data;
+```
+
+Omitting `replaceSectionResultMode` uses the server's `full_song` default:
+two complete songs, with no further `concat`. Listen to both results and check
+`audio_url`, `duration`, new words and transitions. Success does not guarantee
+word-for-word lyrics or the original duration; the `duration` request field
+cannot lock a replacement's length. Returned `lyric` is metadata, not a
+transcription, and may be empty if source lyric timing is unavailable.
+
+For another edit, use the selected complete song's `id` and recheck its timing;
+do not reuse old coordinates or send its whole returned lyric as local input.
+For manual selection, request `replaceSectionResultMode: 'candidates'`, then
+call `await client.suno.generate({ action: 'concat', audioId: selectedCandidateId })`
+and wait for the complete song. Candidates are unstitched clips, possibly with
+context. If results still sing old words or grow too long, check the source,
+interval and local lyrics, then the other result; more polling cannot change
+completed audio. Keep the task ID, source ID, request and result audio for support.
+
+### Processing charges
+
+Standard successful-request charges are **0.56 Credits** for `action: 'stems'`
+(vocals and accompaniment together count once), **0.08 Credits** for
+`client.suno.lyrics(...)` (multiple candidates still count once), and **0.15 Credits**
+for `client.suno.voices(...)`. Failed requests are not charged; applicable discounts
+follow the actual bill. `all_stems`, `extract_stem` and other actions have their own
+prices, not the `stems` rate.
+
 `await client.suno.persona({ name, audioId })` creates a singer style with
 `POST /suno/persona`; creation still follows the endpoint's pricing rules.
 Listing with `GET /suno/persona` and deleting with `DELETE /suno/persona` are

package/dist/index.js

@@ -2266,8 +2266,9 @@ var Suno = class {
     if (options.replaceSectionEnd !== void 0) body["replace_section_end"] = options.replaceSectionEnd;
     if (options.underpaintingStart !== void 0) body["underpainting_start"] = options.underpaintingStart;
     if (options.replaceSectionStart !== void 0) body["replace_section_start"] = options.replaceSectionStart;
+    if (options.replaceSectionResultMode !== void 0) body["replace_section_result_mode"] = options.replaceSectionResultMode;
     for (const [key, value] of Object.entries(options)) {
-      if (!["action", "async", "audioId", "audioUrls", "audioWeight", "callbackUrl", "continueAt", "custom", "duration", "instrumental", "lyric", "lyricPrompt", "mashupAudioIds", "maxWait", "model", "negativeTags", "overpaintingEnd", "overpaintingStart", "personaId", "pollInterval", "prompt", "replaceSectionEnd", "replaceSectionStart", "samplesEnd", "samplesStart", "style", "styleInfluence", "title", "underpaintingEnd", "underpaintingStart", "variationCategory", "vocalGender", "wait", "weirdness"].includes(key) && value !== void 0) {
+      if (!["action", "async", "audioId", "audioUrls", "audioWeight", "callbackUrl", "continueAt", "custom", "duration", "instrumental", "lyric", "lyricPrompt", "mashupAudioIds", "maxWait", "model", "negativeTags", "overpaintingEnd", "overpaintingStart", "personaId", "pollInterval", "prompt", "replaceSectionEnd", "replaceSectionResultMode", "replaceSectionStart", "samplesEnd", "samplesStart", "style", "styleInfluence", "title", "underpaintingEnd", "underpaintingStart", "variationCategory", "vocalGender", "wait", "weirdness"].includes(key) && value !== void 0) {
         body[key] = value;
       }
     }