[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8,"_9":10},"actionData","errors","root",{"_153":154,"_155":156},"routes/_app.x.$handle.$name",{"_11":12,"_13":14,"_91":-7,"_92":43,"_29":-7,"_93":-5,"_15":94,"_95":96,"_97":98,"_99":100,"_101":102,"_103":-5,"_104":-5,"_105":-5,"_106":38,"_107":108,"_109":110,"_111":-5,"_112":-5,"_113":114},"routes/_app.x.$handle.$name.branches",{"_11":12,"_13":14,"_15":16,"_17":18},"handle","rafawiktorowicz","name","Elevenlabs","project",{"_20":28,"_13":14,"_22":36,"_35":-5,"_37":38,"_39":40,"_41":40,"_42":43,"_44":-5,"_45":46,"_47":-5,"_48":38,"_49":-5,"_50":-5,"_51":52,"_53":54,"_55":56,"_57":58,"_59":60,"_61":-5,"_62":63,"_64":63,"_65":66,"_67":66,"_68":66,"_69":63,"_70":63,"_71":72,"_73":74,"_75":76,"_77":38},"branches",[19],{"_20":21,"_22":23,"_24":25,"_13":26,"_27":28,"_29":30,"_31":-5,"_32":-5,"_33":34,"_35":-5},"id","81ac8dbc-5954-11f1-a20a-ee650bb23af1","createdAt",["D",1779835491288],"updatedAt",["D",1780118306455],"main","projectId","81abf186-5954-11f1-a20a-ee650bb23af1","version",21,"forkedBranchId","forkedBranchVersion","lockfileKey","1780118306997-esm.town_v_rafawiktorowicz_Elevenlabs@21-main","deletedAt",["D",1779835491284],"legacyEnvVars",false,"privacy","public","httpPrivacy","description","","imageUrl","userId","52376bcb-d641-4952-81c2-24aad07615a6","pinnedFileId","pinPreview","pinMigrated","screenshotVersion","likeCount",0,"orgId","019e5fc1-0ea2-72fa-904c-6707220b8246","tursoName","81abf186595411f1a20aee650bb23af1","tursoHostname","81abf186595411f1a20aee650bb23af1-val-town.aws-us-east-2.turso.io","tursoMeasuredAt",["D",1789509672838],"pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm","user",{"_11":12,"_90":86},"org",{"_11":12,"_83":84,"_85":86,"_45":46,"_87":88,"_89":38},"branch",{"_29":30,"_13":26,"_20":21,"_27":28,"_31":-5,"_32":-5,"_78":-5,"_79":-5,"_80":-5,"_81":14,"_82":12},"isOwned","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","displayName","Rafał Wiktorowicz","avatar","https://lh3.googleusercontent.com/a/ACg8ocKbp75cCjXsr6hEF5LeT4ePV1zdflGOWHOIG9GAIuLH5-GNTg=s96-c","tier","free","canRestrictVals","profileImageUrl","branchName","path","adoptOrgId",{"_20":28,"_13":14,"_22":148,"_35":-5,"_37":38,"_39":40,"_41":40,"_42":43,"_44":-5,"_45":46,"_47":-5,"_48":38,"_49":-5,"_50":-5,"_51":52,"_53":54,"_55":56,"_57":58,"_59":149,"_61":-5,"_62":63,"_64":63,"_65":66,"_67":66,"_68":66,"_69":63,"_70":63,"_71":150,"_73":151,"_75":152,"_77":38},"tree",[143],"files",[143],"counts",{"_133":52,"_17":134,"_135":134,"_136":137,"_138":139},"contents",[115],"file","upstreamProject","upstreamBranch","emptyEnvVars","flags",[],"hasHttpFiles",true,"grants","orgs","customDomains",{},{"_20":116,"_13":117,"_118":119,"_120":-5,"_121":21,"_122":30,"_123":110,"_22":124,"_29":125,"_126":127,"_128":129,"_130":38,"_131":-5,"_92":132},"fca91442-5be6-11f1-9d3f-ee650bb23af1","main.ts","fileId","81afddd2-5954-11f1-a20a-ee650bb23af1","parentId","branchId","branchVersion","latest",["D",1780118306455],20,"fileType","http","content","// ============================================================================\n// ElevenLabs MCP connector — v10 (OPTYMALNA: mądry podział po akapitach + wielowątek + retry + job bez timeoutu; v3-safe, BEZ previous_text/next_text)\n// speed top-level FIX w v6.1+\n// ============================================================================\nimport { blob } from \"https://esm.town/v/std/blob\";\nconst KEY = Deno.env.get(\"ELEVENLABS_API_KEY\") ?? \"\";\nconst EL = \"https://api.elevenlabs.io\";\nlet SELF_ORIGIN = \"\";\nconst MODEL_LIMITS: Record = {\n eleven_multilingual_v2: 9000,\n eleven_turbo_v2_5: 38000,\n eleven_flash_v2_5: 38000,\n eleven_v3: 4500,\n};\nfunction stripId3(b: Uint8Array): Uint8Array {\n let s = 0, e = b.length;\n if (b.length > 10 && b[0] === 0x49 && b[1] === 0x44 && b[2] === 0x33) {\n s = 10 +\n (((b[6] & 0x7f) << 21) | ((b[7] & 0x7f) << 14) | ((b[8] & 0x7f) << 7) |\n (b[9] & 0x7f));\n }\n if (\n e - s >= 128 && b[e - 128] === 0x54 && b[e - 127] === 0x41 &&\n b[e - 126] === 0x47\n ) e -= 128;\n return b.subarray(s, e);\n}\nfunction frameLen(b: Uint8Array, i: number): number {\n if (b[i] !== 0xFF || (b[i + 1] & 0xE0) !== 0xE0) return 0;\n const ver = (b[i + 1] >> 3) & 3, layer = (b[i + 1] >> 1) & 3;\n if (layer !== 1) return 0;\n const brIdx = (b[i + 2] >> 4) & 0xF,\n srIdx = (b[i + 2] >> 2) & 3,\n pad = (b[i + 2] >> 1) & 1;\n const BR1 = [\n 0,\n 32,\n 40,\n 48,\n 56,\n 64,\n 80,\n 96,\n 112,\n 128,\n 160,\n 192,\n 224,\n 256,\n 320,\n 0,\n ];\n const BR2 = [0, 8, 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128, 144, 160, 0];\n const SR1 = [44100, 48000, 32000, 0],\n SR2 = [22050, 24000, 16000, 0],\n SR25 = [11025, 12000, 8000, 0];\n let br = 0, sr = 0, spf = 1152;\n if (ver === 3) {\n br = BR1[brIdx] * 1000;\n sr = SR1[srIdx];\n spf = 1152;\n } else if (ver === 2) {\n br = BR2[brIdx] * 1000;\n sr = SR2[srIdx];\n spf = 576;\n } else if (ver === 0) {\n br = BR2[brIdx] * 1000;\n sr = SR25[srIdx];\n spf = 576;\n }\n if (!br || !sr) return 0;\n return Math.floor((spf / 8) * br / sr) + pad;\n}\nfunction stripXing(b: Uint8Array): Uint8Array {\n let i = 0;\n while (i < b.length - 4 && !(b[i] === 0xFF && (b[i + 1] & 0xE0) === 0xE0)) {\n i++;\n }\n if (i >= b.length - 4) return b;\n const len = frameLen(b, i);\n if (len > 0 && i + len <= b.length) {\n let x = false;\n for (let k = i + 4; k < i + Math.min(len, 60) - 4; k++) {\n if (\n (b[k] === 0x58 && b[k + 1] === 0x69 && b[k + 2] === 0x6E &&\n b[k + 3] === 0x67) ||\n (b[k] === 0x49 && b[k + 1] === 0x6E && b[k + 2] === 0x66 &&\n b[k + 3] === 0x6F)\n ) {\n x = true;\n break;\n }\n }\n if (x) return b.subarray(i + len);\n }\n return b.subarray(i);\n}\nfunction cleanMp3(b: Uint8Array): Uint8Array {\n return stripXing(stripId3(b));\n}\nfunction concat(parts: Uint8Array[]): Uint8Array {\n const total = parts.reduce((n, p) => n + p.length, 0);\n const out = new Uint8Array(total);\n let off = 0;\n for (const p of parts) {\n out.set(p, off);\n off += p.length;\n }\n return out;\n}\nconst TOOLS = [\n {\n name: \"tts\",\n description:\n \"Generuje mowę z tekstu głosem ElevenLabs i zwraca PUBLICZNY link do MP3.\",\n inputSchema: {\n type: \"object\",\n properties: {\n text: { type: \"string\" },\n voice_id: { type: \"string\" },\n model_id: { type: \"string\", default: \"eleven_multilingual_v2\" },\n stability: { type: \"number\", default: 0.5 },\n similarity_boost: { type: \"number\", default: 0.8 },\n style: { type: \"number\", default: 0.0 },\n use_speaker_boost: { type: \"boolean\", default: true },\n speed: {\n type: \"number\",\n description: \"0.7–1.2. 0.9=10% wolniej.\",\n default: 1.0,\n },\n },\n required: [\"text\", \"voice_id\"],\n },\n },\n {\n name: \"tts_batch\",\n description:\n \"★ SZYBKIE: generuje WIELE kawałków równolegle (Promise.all). Zamiast N osobnych tts — JEDNO wywołanie.\",\n inputSchema: {\n type: \"object\",\n properties: {\n texts: {\n type: \"array\",\n items: { type: \"string\" },\n description: \"Lista tekstów ~600–750 zn.\",\n },\n voice_id: { type: \"string\" },\n model_id: { type: \"string\", default: \"eleven_v3\" },\n stability: { type: \"number\", default: 0.5 },\n similarity_boost: { type: \"number\", default: 0.85 },\n style: { type: \"number\", default: 0.0 },\n use_speaker_boost: { type: \"boolean\", default: true },\n speed: { type: \"number\", default: 0.9 },\n max_concurrent: { type: \"number\", default: 5 },\n },\n required: [\"texts\", \"voice_id\"],\n },\n },\n {\n name: \"tts_longform\",\n description:\n \"AUDIOBOOK: dzieli długi tekst, generuje kawałki sekwencyjnie i skleja.\",\n inputSchema: {\n type: \"object\",\n properties: {\n text: { type: \"string\" },\n voice_id: { type: \"string\" },\n model_id: { type: \"string\", default: \"eleven_multilingual_v2\" },\n stability: { type: \"number\", default: 0.6 },\n similarity_boost: { type: \"number\", default: 0.85 },\n style: { type: \"number\", default: 0.0 },\n use_speaker_boost: { type: \"boolean\", default: true },\n speed: { type: \"number\", default: 1.0 },\n },\n required: [\"text\", \"voice_id\"],\n },\n },\n {\n name: \"stitch_audio\",\n description:\n \"Skleja pliki (po ID z tts/tts_batch) w JEDEN plik MP3. 'gap:2000' = 2s ciszy.\",\n inputSchema: {\n type: \"object\",\n properties: {\n ids: { type: \"array\", items: { type: \"string\" } },\n gap_ms: { type: \"number\", default: 1500 },\n },\n required: [\"ids\"],\n },\n },\n {\n name: \"list_voices\",\n description: \"Lista głosów na koncie ElevenLabs.\",\n inputSchema: { type: \"object\", properties: {} },\n },\n {\n name: \"search_library\",\n description: \"Przeszukuje bibliotekę głosów ElevenLabs.\",\n inputSchema: {\n type: \"object\",\n properties: {\n language: { type: \"string\", default: \"pl\" },\n gender: { type: \"string\" },\n age: { type: \"string\" },\n search: { type: \"string\" },\n page_size: { type: \"number\", default: 20 },\n },\n },\n },\n {\n name: \"voice_design\",\n description: \"Projektuje NOWY głos z opisu i zapisuje na koncie.\",\n inputSchema: {\n type: \"object\",\n properties: {\n prompt: { type: \"string\" },\n sample_text: { type: \"string\" },\n name: { type: \"string\", default: \"Starsza Polka (design)\" },\n loudness: { type: \"number\", default: 0.5 },\n guidance_scale: { type: \"number\", default: 30 },\n model_id: { type: \"string\", default: \"eleven_multilingual_ttv_v2\" },\n auto_create: { type: \"boolean\", default: true },\n },\n required: [\"prompt\", \"sample_text\"],\n },\n },\n {\n name: \"remix_voice\",\n description: \"Remiksuje istniejący głos nowym promptem.\",\n inputSchema: {\n type: \"object\",\n properties: {\n voice_id: { type: \"string\" },\n prompt: { type: \"string\" },\n sample_text: { type: \"string\" },\n },\n required: [\"voice_id\", \"prompt\", \"sample_text\"],\n },\n },\n {\n name: \"tts_audiobook\",\n description:\n \"★ AUDIOBOOK 1 wywołanie (krótkie/średnie): mądry podział po akapitach (przejścia w naturalnych pauzach), generuje RÓWNOLEGLE v3 z retry i SKLEJA w jeden plik.\",\n inputSchema: {\n type: \"object\",\n properties: {\n text: { type: \"string\" },\n voice_id: { type: \"string\" },\n model_id: { type: \"string\", default: \"eleven_v3\" },\n stability: { type: \"number\", default: 0.6 },\n similarity_boost: { type: \"number\", default: 0.85 },\n style: { type: \"number\", default: 0.0 },\n use_speaker_boost: { type: \"boolean\", default: true },\n speed: { type: \"number\", default: 0.9 },\n target_chars: { type: \"number\", default: 2200 },\n max_concurrent: { type: \"number\", default: 5 },\n gap_ms: { type: \"number\", default: 350 },\n },\n required: [\"text\", \"voice_id\"],\n },\n },\n {\n name: \"audiobook_start\",\n description:\n \"★ GODZINY, wielowątkowo, BEZ timeoutu: zwraca {job_id} od razu. Potem odpytuj audiobook_get — kazde wywolanie robi jedna fale rownolegla (batch_size) i wraca. Madry podzial po akapitach.\",\n inputSchema: {\n type: \"object\",\n properties: {\n text: { type: \"string\" },\n voice_id: { type: \"string\" },\n model_id: { type: \"string\", default: \"eleven_v3\" },\n stability: { type: \"number\", default: 0.6 },\n similarity_boost: { type: \"number\", default: 0.85 },\n style: { type: \"number\", default: 0.0 },\n use_speaker_boost: { type: \"boolean\", default: true },\n speed: { type: \"number\", default: 0.9 },\n target_chars: { type: \"number\", default: 2200 },\n max_concurrent: { type: \"number\", default: 5 },\n gap_ms: { type: \"number\", default: 350 },\n batch_size: { type: \"number\", default: 5 },\n },\n required: [\"text\", \"voice_id\"],\n },\n },\n {\n name: \"audiobook_get\",\n description: \"Status zadania audiobook: {status,total,audio_url?}.\",\n inputSchema: {\n type: \"object\",\n properties: { job_id: { type: \"string\" } },\n required: [\"job_id\"],\n },\n },\n {\n name: \"get_audio\",\n description: \"Zwraca linki do kawałków po ID.\",\n inputSchema: {\n type: \"object\",\n properties: { ids: { type: \"array\", items: { type: \"string\" } } },\n required: [\"ids\"],\n },\n },\n];\nfunction chunkText(text: string, target: number): string[] {\n const paras = text.replace(/\\r/g, \"\").split(/\\n\\s*\\n+/).map((p) => p.trim())\n .filter(Boolean);\n const sentSplit = (p: string) =>\n p.match(/[^.!?…]+[.!?…]+[\"”»)\\]]*\\s*|[^.!?…]+$/g) ?? [p];\n const out: string[] = [];\n let cur = \"\";\n const push = () => {\n if (cur.trim()) {\n out.push(cur.trim());\n cur = \"\";\n }\n };\n for (const para of paras) {\n if ((cur ? cur.length + 2 : 0) + para.length <= target) {\n cur = cur ? cur + \"\\n\\n\" + para : para;\n continue;\n }\n if (cur) push();\n if (para.length <= target) {\n cur = para;\n continue;\n }\n for (const s of sentSplit(para)) {\n if (cur && (cur + s).length > target) push();\n if (s.length > target) {\n for (const w of s.split(/\\s+/)) {\n if (cur && (cur + \" \" + w).length > target) push();\n cur += (cur ? \" \" : \"\") + w;\n }\n } else cur += s;\n }\n push();\n }\n push();\n return out;\n}\nfunction vs(a: Record, sd: number, sm: number) {\n return {\n stability: Number(a.stability ?? sd),\n similarity_boost: Number(a.similarity_boost ?? sm),\n style: Number(a.style ?? 0),\n use_speaker_boost: a.use_speaker_boost !== false,\n speed: Number(a.speed ?? 1.0),\n };\n}\nasync function ttsBytes(\n text: string,\n voice_id: unknown,\n model_id: string,\n settings: object,\n): Promise {\n const { speed, ...voiceSettings } = settings as any;\n const body = JSON.stringify({\n text,\n model_id,\n voice_settings: voiceSettings,\n speed,\n });\n for (let att = 0;; att++) {\n const r = await fetch(\n `${EL}/v1/text-to-speech/${voice_id}?output_format=mp3_44100_128`,\n {\n method: \"POST\",\n headers: { \"xi-api-key\": KEY, \"content-type\": \"application/json\" },\n body,\n },\n );\n if (r.ok) return new Uint8Array(await r.arrayBuffer());\n if ((r.status === 429 || r.status >= 500) && att < 5) {\n await new Promise((z) =>\n setTimeout(z, 800 * 2 ** att + Math.random() * 400)\n );\n continue;\n }\n throw new Error(`TTS ${r.status}: ${await r.text()}`);\n }\n}\nasync function loadAudio(id: string): Promise {\n const b = await blob.get(`tts_${id}`);\n if (!b) throw new Error(`brak audio ${id}`);\n return cleanMp3(new Uint8Array(await b.arrayBuffer()));\n}\nasync function doTts(a: Record) {\n const bytes = await ttsBytes(\n String(a.text ?? \"\"),\n a.voice_id,\n String(a.model_id ?? \"eleven_multilingual_v2\"),\n vs(a, 0.5, 0.8),\n );\n const id = crypto.randomUUID();\n await blob.set(`tts_${id}`, new Blob([bytes], { type: \"audio/mpeg\" }));\n return {\n audio_url: `${SELF_ORIGIN}/audio/${id}`,\n id,\n bytes: bytes.length,\n voice_id: a.voice_id,\n };\n}\nasync function doTtsBatch(a: Record) {\n const texts = (a.texts as string[]) ?? [];\n if (!texts.length) throw new Error(\"Brak tekstów.\");\n const model_id = String(a.model_id ?? \"eleven_v3\");\n const settings = vs(a, 0.5, 0.85);\n const maxC = Math.min(Number(a.max_concurrent ?? 5), texts.length);\n const results: Array<{ id: string; audio_url: string; bytes: number }> =\n new Array(texts.length);\n const queue = texts.map((text, idx) => ({ text, idx }));\n const worker = async () => {\n while (queue.length > 0) {\n const item = queue.shift();\n if (!item) break;\n let bytes: Uint8Array;\n try {\n bytes = await ttsBytes(item.text, a.voice_id, model_id, settings);\n } catch (e) {\n await new Promise((r) => setTimeout(r, 1000));\n bytes = await ttsBytes(item.text, a.voice_id, model_id, settings);\n }\n const id = crypto.randomUUID();\n await blob.set(`tts_${id}`, new Blob([bytes], { type: \"audio/mpeg\" }));\n results[item.idx] = {\n id,\n audio_url: `${SELF_ORIGIN}/audio/${id}`,\n bytes: bytes.length,\n };\n }\n };\n await Promise.all(Array(maxC).fill(null).map(() => worker()));\n return {\n ids: results.map((r) => r.id),\n audio_urls: results.map((r) => r.audio_url),\n count: texts.length,\n bytes_total: results.reduce((n, r) => n + r.bytes, 0),\n };\n}\nasync function doTtsLongform(a: Record) {\n const model_id = String(a.model_id ?? \"eleven_multilingual_v2\");\n const text = String(a.text ?? \"\");\n const chunks = chunkText(text, MODEL_LIMITS[model_id] ?? 9000);\n if (!chunks.length) throw new Error(\"Brak tekstu.\");\n const parts: Uint8Array[] = [];\n for (const c of chunks) {\n parts.push(\n cleanMp3(await ttsBytes(c, a.voice_id, model_id, vs(a, 0.6, 0.85))),\n );\n await new Promise((r) => setTimeout(r, 150));\n }\n const merged = concat(parts);\n const id = crypto.randomUUID();\n await blob.set(`tts_${id}`, new Blob([merged], { type: \"audio/mpeg\" }));\n return {\n audio_url: `${SELF_ORIGIN}/audio/${id}`,\n id,\n chunks: chunks.length,\n chars: text.length,\n bytes: merged.length,\n seconds_est: Math.round(text.length / 15),\n voice_id: a.voice_id,\n model_id,\n };\n}\nasync function doStitch(a: Record) {\n const ids = (a.ids as string[]) ?? [];\n if (!ids.length) throw new Error(\"Brak ID.\");\n const defGap = Number(a.gap_ms ?? 1500);\n const cache: Record = {};\n let tpl: { header: Uint8Array; len: number } | null = null;\n for (const raw of ids) {\n const s = String(raw).trim();\n if (/^gap(?::\\d+)?$/i.test(s)) continue;\n const id = s.split(\"/\").pop()!;\n if (cache[id]) continue;\n const b = await blob.get(`tts_${id}`);\n if (!b) throw new Error(`Nie znaleziono: ${id}`);\n const clean = cleanMp3(new Uint8Array(await b.arrayBuffer()));\n cache[id] = clean;\n if (!tpl) {\n let i = 0;\n while (\n i < clean.length - 4 &&\n !(clean[i] === 0xFF && (clean[i + 1] & 0xE0) === 0xE0)\n ) i++;\n const len = frameLen(clean, i);\n if (len > 0) tpl = { header: clean.slice(i, i + 4), len };\n }\n }\n const sil = (ms: number): Uint8Array => {\n if (!tpl) return new Uint8Array(0);\n const fMs = 1152 / 44100 * 1000;\n const n = Math.max(1, Math.round(ms / fMs));\n const out = new Uint8Array(tpl.len * n);\n for (let k = 0; k < n; k++) out.set(tpl.header, k * tpl.len);\n return out;\n };\n const parts: Uint8Array[] = [];\n for (const raw of ids) {\n const s = String(raw).trim();\n const m = s.match(/^gap(?::(\\d+))?$/i);\n if (m) {\n parts.push(sil(m[1] ? Number(m[1]) : defGap));\n continue;\n }\n parts.push(cache[s.split(\"/\").pop()!]);\n }\n const merged = concat(parts);\n const id = crypto.randomUUID();\n await blob.set(`tts_${id}`, new Blob([merged], { type: \"audio/mpeg\" }));\n return {\n audio_url: `${SELF_ORIGIN}/audio/${id}`,\n id,\n parts: ids.length,\n bytes: merged.length,\n };\n}\nasync function doListVoices() {\n const r = await fetch(`${EL}/v1/voices`, { headers: { \"xi-api-key\": KEY } });\n if (!r.ok) throw new Error(`voices ${r.status}`);\n const j = await r.json();\n return (j.voices ?? []).map((v: any) => ({\n voice_id: v.voice_id,\n name: v.name,\n labels: v.labels ?? {},\n category: v.category,\n }));\n}\nasync function doSearchLibrary(a: Record) {\n const p = new URLSearchParams();\n if (a.language) p.set(\"language\", String(a.language));\n if (a.gender) p.set(\"gender\", String(a.gender));\n if (a.age) p.set(\"age\", String(a.age));\n if (a.search) p.set(\"search\", String(a.search));\n p.set(\"page_size\", String(a.page_size ?? 20));\n const r = await fetch(`${EL}/v1/shared-voices?${p}`, {\n headers: { \"xi-api-key\": KEY },\n });\n if (!r.ok) throw new Error(`${r.status}`);\n const j = await r.json();\n return (j.voices ?? []).map((v: any) => ({\n voice_id: v.voice_id,\n name: v.name,\n gender: v.gender,\n age: v.age,\n language: v.language,\n preview_url: v.preview_url,\n }));\n}\nasync function doVoiceDesign(a: Record) {\n const d = await fetch(`${EL}/v1/text-to-voice/design`, {\n method: \"POST\",\n headers: { \"xi-api-key\": KEY, \"content-type\": \"application/json\" },\n body: JSON.stringify({\n voice_description: String(a.prompt ?? \"\"),\n text: String(a.sample_text ?? \"\"),\n model_id: String(a.model_id ?? \"eleven_multilingual_ttv_v2\"),\n loudness: Number(a.loudness ?? 0.5),\n guidance_scale: Number(a.guidance_scale ?? 30),\n }),\n });\n if (!d.ok) throw new Error(`design ${d.status}: ${await d.text()}`);\n const dj = await d.json();\n const previews = (dj.previews ?? []).map((p: any) => ({\n generated_voice_id: p.generated_voice_id,\n }));\n if (!a.auto_create || !previews.length) return { previews };\n const c = await fetch(`${EL}/v1/text-to-voice`, {\n method: \"POST\",\n headers: { \"xi-api-key\": KEY, \"content-type\": \"application/json\" },\n body: JSON.stringify({\n voice_name: String(a.name ?? \"Starsza Polka\"),\n voice_description: String(a.prompt ?? \"\"),\n generated_voice_id: previews[0].generated_voice_id,\n }),\n });\n if (!c.ok) throw new Error(`create-voice ${c.status}`);\n const cj = await c.json();\n return {\n voice_id: cj.voice_id ?? cj.voice?.voice_id,\n name: String(a.name ?? \"\"),\n previews_count: previews.length,\n };\n}\nasync function doRemix(a: Record) {\n const r = await fetch(`${EL}/v1/text-to-voice/${a.voice_id}/remix`, {\n method: \"POST\",\n headers: { \"xi-api-key\": KEY, \"content-type\": \"application/json\" },\n body: JSON.stringify({\n voice_description: String(a.prompt ?? \"\"),\n text: String(a.sample_text ?? \"\"),\n }),\n });\n if (!r.ok) throw new Error(`remix ${r.status}`);\n const j = await r.json();\n return {\n previews: (j.previews ?? []).map((p: any) => ({\n generated_voice_id: p.generated_voice_id,\n })),\n };\n}\nfunction buildSilence(sample: Uint8Array, ms: number): Uint8Array {\n let i = 0;\n while (\n i < sample.length - 4 &&\n !(sample[i] === 0xFF && (sample[i + 1] & 0xE0) === 0xE0)\n ) i++;\n const len = frameLen(sample, i);\n if (len <= 0 || ms <= 0) return new Uint8Array(0);\n const header = sample.slice(i, i + 4);\n const fMs = 1152 / 44100 * 1000;\n const n = Math.max(1, Math.round(ms / fMs));\n const out = new Uint8Array(len * n);\n for (let k = 0; k < n; k++) out.set(header, k * len);\n return out;\n}\nasync function genParallel(\n chunks: string[],\n voice_id: unknown,\n model_id: string,\n settings: object,\n maxC: number,\n idxs?: number[],\n): Promise<{ parts: Record; failed: number[] }> {\n const targets = idxs ?? chunks.map((_, i) => i);\n const parts: Record = {};\n const failed: number[] = [];\n const queue = [...targets];\n const worker = async () => {\n while (queue.length) {\n const i = queue.shift();\n if (i === undefined) break;\n try {\n parts[i] = cleanMp3(\n await ttsBytes(chunks[i], voice_id, model_id, settings),\n );\n } catch (_e) {\n failed.push(i);\n }\n }\n };\n await Promise.all(\n Array(Math.max(1, Math.min(maxC, targets.length))).fill(null).map(() =>\n worker()\n ),\n );\n return { parts, failed };\n}\nasync function buildAudiobook(a: Record) {\n const model_id = String(a.model_id ?? \"eleven_v3\");\n const text = String(a.text ?? \"\");\n const target = Math.min(\n Number(a.target_chars ?? 2200),\n MODEL_LIMITS[model_id] ?? 4500,\n );\n const chunks = chunkText(text, target);\n if (!chunks.length) throw new Error(\"Brak tekstu.\");\n const settings = vs(a, 0.6, 0.85);\n const maxC = Number(a.max_concurrent ?? 5);\n const gap = Number(a.gap_ms ?? 350);\n const { parts, failed } = await genParallel(\n chunks,\n a.voice_id,\n model_id,\n settings,\n maxC,\n );\n if (failed.length) {\n return {\n error: \"część kawałków padła — ponów\",\n failed,\n chunks: chunks.length,\n };\n }\n const arr = chunks.map((_, i) => parts[i]);\n const sample = arr.find((p) => p && p.length)!;\n const sil = buildSilence(sample, gap);\n const merged: Uint8Array[] = [];\n for (let i = 0; i < arr.length; i++) {\n merged.push(arr[i]);\n if (gap > 0 && i < arr.length - 1) merged.push(sil);\n }\n const out = concat(merged);\n const id = crypto.randomUUID();\n await blob.set(`tts_${id}`, new Blob([out], { type: \"audio/mpeg\" }));\n return {\n audio_url: `${SELF_ORIGIN}/audio/${id}`,\n id,\n chunks: chunks.length,\n bytes: out.length,\n chars: text.length,\n seconds_est: Math.round(text.length / 15),\n voice_id: a.voice_id,\n model_id,\n };\n}\nasync function doTtsAudiobook(a: Record) {\n return await buildAudiobook(a);\n}\nasync function doAudiobookStart(a: Record) {\n const model_id = String(a.model_id ?? \"eleven_v3\");\n const text = String(a.text ?? \"\");\n const target = Math.min(\n Number(a.target_chars ?? 2200),\n MODEL_LIMITS[model_id] ?? 4500,\n );\n const chunks = chunkText(text, target);\n if (!chunks.length) throw new Error(\"Brak tekstu.\");\n const maxC = Number(a.max_concurrent ?? 5);\n const job_id = crypto.randomUUID();\n const state = {\n status: \"processing\",\n total: chunks.length,\n done: 0,\n chunks,\n ids: new Array(chunks.length).fill(null),\n voice_id: a.voice_id,\n model_id,\n settings: vs(a, 0.6, 0.85),\n gap_ms: Number(a.gap_ms ?? 350),\n batch_size: Number(a.batch_size ?? maxC),\n max_concurrent: maxC,\n lock: 0,\n };\n await blob.set(\n `job_${job_id}`,\n new Blob([JSON.stringify(state)], { type: \"application/json\" }),\n );\n return {\n job_id,\n total: chunks.length,\n status: \"processing\",\n hint:\n \"Odpytuj audiobook_get(job_id) az status=done; kazde wywolanie robi jedna fale rownolegla — bez timeoutu.\",\n };\n}\nasync function doAudiobookGet(a: Record) {\n const key = `job_${String(a.job_id)}`;\n let b: any;\n try {\n b = await blob.get(key);\n } catch {\n return { status: \"not_found\" };\n }\n if (!b) return { status: \"not_found\" };\n const s: any = JSON.parse(await b.text());\n if (s.status === \"done\" || s.status === \"error\") {\n return {\n status: s.status,\n done: s.done,\n total: s.total,\n audio_url: s.audio_url,\n error: s.error,\n };\n }\n const now = Date.now();\n if (s.lock && now - s.lock < 90000) {\n return {\n status: \"processing\",\n done: s.done,\n total: s.total,\n note: \"partia w toku\",\n };\n }\n s.lock = now;\n await blob.set(\n key,\n new Blob([JSON.stringify(s)], { type: \"application/json\" }),\n );\n try {\n const pending: number[] = [];\n for (let i = 0; i < s.chunks.length && pending.length < s.batch_size; i++) {\n if (!s.ids[i]) pending.push(i);\n }\n const { parts, failed } = await genParallel(\n s.chunks,\n s.voice_id,\n s.model_id,\n s.settings,\n s.max_concurrent,\n pending,\n );\n for (const i of pending) {\n if (parts[i]) {\n const nid = crypto.randomUUID();\n await blob.set(\n `tts_${nid}`,\n new Blob([parts[i]], { type: \"audio/mpeg\" }),\n );\n s.ids[i] = nid;\n }\n }\n s.done = s.ids.filter(Boolean).length;\n s.lock = 0;\n if (s.done >= s.total) {\n const sample = await loadAudio(s.ids.find((x: any) => x));\n const sil = buildSilence(sample, s.gap_ms);\n const arr: Uint8Array[] = [];\n for (let i = 0; i < s.ids.length; i++) {\n arr.push(await loadAudio(s.ids[i]));\n if (s.gap_ms > 0 && i < s.ids.length - 1) arr.push(sil);\n }\n const out = concat(arr);\n const fid = crypto.randomUUID();\n await blob.set(`tts_${fid}`, new Blob([out], { type: \"audio/mpeg\" }));\n s.status = \"done\";\n s.audio_url = `${SELF_ORIGIN}/audio/${fid}`;\n s.bytes = out.length;\n }\n await blob.set(\n key,\n new Blob([JSON.stringify(s)], { type: \"application/json\" }),\n );\n return {\n status: s.status,\n done: s.done,\n total: s.total,\n failed_now: failed,\n audio_url: s.audio_url,\n };\n } catch (e) {\n s.lock = 0;\n await blob.set(\n key,\n new Blob([JSON.stringify(s)], { type: \"application/json\" }),\n );\n return {\n status: \"processing\",\n done: s.done,\n total: s.total,\n error_last: String(e).slice(0, 160),\n };\n }\n}\nasync function doGetAudio(a: Record) {\n const ids = (a.ids as string[]) ?? [];\n return {\n audio_urls: ids.map((id) =>\n `${SELF_ORIGIN}/audio/${String(id).split(\"/\").pop()}`\n ),\n count: ids.length,\n };\n}\nasync function callTool(name: string, args: Record) {\n if (name === \"tts\") return await doTts(args);\n if (name === \"tts_batch\") return await doTtsBatch(args);\n if (name === \"tts_longform\") return await doTtsLongform(args);\n if (name === \"stitch_audio\") return await doStitch(args);\n if (name === \"list_voices\") return await doListVoices();\n if (name === \"search_library\") return await doSearchLibrary(args);\n if (name === \"voice_design\") return await doVoiceDesign(args);\n if (name === \"remix_voice\") return await doRemix(args);\n if (name === \"tts_audiobook\") return await doTtsAudiobook(args);\n if (name === \"audiobook_start\") return await doAudiobookStart(args);\n if (name === \"audiobook_get\") return await doAudiobookGet(args);\n if (name === \"get_audio\") return await doGetAudio(args);\n throw new Error(`Unknown tool: ${name}`);\n}\nfunction rpc(id: unknown, result?: unknown, error?: unknown) {\n const body: Record = { jsonrpc: \"2.0\", id };\n if (error) body.error = error;\n else body.result = result;\n return new Response(JSON.stringify(body), {\n headers: { \"content-type\": \"application/json\" },\n });\n}\nexport default async function (req: Request): Promise {\n const url = new URL(req.url);\n SELF_ORIGIN = url.origin;\n if (req.method === \"GET\" && url.pathname.startsWith(\"/audio/\")) {\n const id = url.pathname.slice(\"/audio/\".length);\n try {\n const b = await blob.get(`tts_${id}`);\n return new Response(await b.arrayBuffer(), {\n headers: { \"content-type\": \"audio/mpeg\" },\n });\n } catch {\n return new Response(\"not found\", { status: 404 });\n }\n }\n if (req.method === \"GET\" && url.pathname === \"/\") {\n return new Response(\"ElevenLabs MCP v10. POST /mcp\");\n }\n if (\n req.method === \"POST\" && (url.pathname === \"/mcp\" || url.pathname === \"/\")\n ) {\n let msg: any;\n try {\n msg = await req.json();\n } catch {\n return rpc(null, undefined, { code: -32700, message: \"Parse error\" });\n }\n const { id, method, params } = msg ?? {};\n if (method === \"initialize\") {\n return rpc(id, {\n protocolVersion: \"2024-11-05\",\n capabilities: { tools: {} },\n serverInfo: { name: \"elevenlabs-mcp\", version: \"10.0.0\" },\n });\n }\n if (method === \"notifications/initialized\") {\n return new Response(null, { status: 204 });\n }\n if (method === \"tools/list\") return rpc(id, { tools: TOOLS });\n if (method === \"tools/call\") {\n try {\n const out = await callTool(params?.name, params?.arguments ?? {});\n return rpc(id, {\n content: [{ type: \"text\", text: JSON.stringify(out, null, 2) }],\n });\n } catch (e) {\n return rpc(id, {\n content: [{ type: \"text\", text: `BŁĄD: ${(e as Error).message}` }],\n isError: true,\n });\n }\n }\n return rpc(id, undefined, {\n code: -32601,\n message: `Method not found: ${method}`,\n });\n }\n return new Response(\"Method not allowed\", { status: 405 });\n}","deleted","label",[117],"pulls",1,"forks","envVars",{"_140":134,"_142":52},"connections",{"_140":52,"_141":52},"count","unfilledCount","emptyCount",{"_20":116,"_13":117,"_118":119,"_120":-5,"_121":21,"_122":30,"_123":110,"_22":144,"_29":125,"_126":127,"_128":129,"_130":38,"_131":-5,"_145":146,"_92":147},["D",1780118306455],"children",[],[117],["D",1779835491284],["D",1789509672838],{"_11":12,"_90":86},{"_11":12,"_83":84,"_85":86,"_45":46,"_87":88,"_89":38},{"_29":30,"_13":26,"_20":21,"_27":28,"_31":-5,"_32":-5,"_78":-5,"_79":-5,"_80":-5,"_81":14,"_82":12},"ctx",{"_71":-5,"_112":186,"_187":188,"_189":190,"_191":38,"_192":38,"_193":38,"_194":110,"_195":-5,"_196":-7,"_197":-7,"_198":52,"_199":38,"_200":38,"_201":-5,"_202":38,"_203":38,"_204":38,"_205":-5},"ENV",{"_157":158,"_159":160,"_161":26,"_162":163,"_164":165,"_166":167,"_168":169,"_170":171,"_172":173,"_174":175,"_176":177,"_178":179,"_180":181,"_182":183,"_184":185},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","RENDER_GIT_COMMIT","e87eb04e707cf64e7e532c8e8ec395f209e17f15","RENDER_GIT_BRANCH","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","GOOGLE_CLIENT_ID","875989696899-jtsqkhlbdv5jrlsk8631p7ipjg3c6o62.apps.googleusercontent.com","STRIPE_PUBLISHABLE_KEY","pk_live_51N2zvWEpZTTLipIkVEwXrUG9AT0yeOOEYxszarZT1jpsZV58Ly6wlg3SltZpHN26CUalOBmUSY7adQRQiSkCZz2500qSXNkf8Y","API_URL","https://api.val.town",[],"preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesCloseBrackets","preferencesChatPanelState","preferencesDprint","preferencesDenofmt","notificationCount","flagDenols","isBusinessOrgMember","businessOrgHandle","isBusinessOrgAdmin","isHigherTownieTier","navCollapsed","currentOrgId"]

Branches1

Branches
1