[{"_1":2,"_170":-5,"_171":-5},"loaderData",{"_3":4,"_13":14,"_21":22,"_144":145},"root",{"_5":-5,"_6":-7,"_7":8,"_9":-7,"_10":11,"_12":-7},"currentUser","primaryEmail","siteVerification","6LeIxAcTAAAAAJcZVRqyHh71UMIEGNQ_MXjiZKhI","toast","enableDev",false,"incidentStatus","routes/_layout",{"_15":-5,"_16":17,"_18":17,"_19":11,"_20":11},"username","isAskGordonEnabled",true,"isUsingNewTemplate","showMinimalLanding","showHoneypotTrap","routes/_layout.r.$namespace.$reponame",{"_23":24,"_25":-5,"_26":11,"_27":11,"_28":29,"_30":31,"_32":-5,"_33":11,"_34":35,"_89":-5,"_90":91,"_135":-5,"_136":11,"_137":17,"_138":11,"_139":-5,"_140":11,"_141":-5,"_142":143},"canonicalUrl","https://hub.docker.com/r/eworkerinc/vibevoice/tags","mcpServer","canSeeDockerfile","canSeeAutobuilds","namespace","eworkerinc","reponame","vibevoice","cloudBuildSource","isAutobuildsAvailable","repo",{"_36":37,"_43":44,"_46":47,"_48":49,"_50":11,"_51":52,"_53":54,"_55":56,"_39":31,"_28":29,"_58":59,"_60":45,"_61":62,"_63":64,"_65":66,"_67":-5,"_68":69,"_70":71,"_72":11,"_73":29,"_74":75,"_80":11,"_81":82,"_86":29,"_87":88},"categories",[38],{"_39":40,"_41":42},"name","Machine learning & AI","slug","machine-learning-and-ai","content_types",[45],"image","date_registered","2025-09-02T20:32:19.379188Z","description","","is_private","last_updated","2025-09-03T18:08:28.795433Z","last_modified","2025-09-05T12:51:35.156325Z","media_types",[57],"application/vnd.docker.container.image.v1+json","pull_count",3399,"repository_type","star_count",7,"status",1,"status_description","active","source","collaborator_count",0,"full_description","# E‑Worker VibeVoice Integration Stack (GPU)\n\n## Overview\n\nThe **E‑Worker VibeVoice Integration Stack (GPU)** is a unified, single‑container deployment that exposes a clean HTTP proxy on one public port while keeping Microsoft’s VibeVoice backends private inside. This stack simplifies setup, enforces security, and enables applications—including **E‑Worker Soundstage**—to integrate seamlessly with high‑quality text‑to‑speech (TTS).\n\nhttps://app.eworker.ca \n\nhttps://www.reddit.com/r/eworker_ca/\n\nhttps://x.com/eworker_ca\n\n**Important: The VibeVoice API are ready to use, the E-Worker Soundstage UI is still under development.**\n\n---\n\n## Important Links and Licensing\n\n* Microsoft VibeVoice code and models:\n\n * Repo: [https://github.com/microsoft/VibeVoice](https://github.com/microsoft/VibeVoice)\n * Model cards: [VibeVoice‑1.5B](https://huggingface.co/microsoft/VibeVoice-1.5B), [VibeVoice‑Large](https://huggingface.co/microsoft/VibeVoice-Large)\n* Licensing: Accept and comply with Microsoft’s repo/model licenses.\n* This image orchestrates upstream VibeVoice without modifying the models.\n* **Disclaimer:** E‑Worker is not affiliated with Microsoft.\n\n---\n\n## What’s Included\n\n* Public port **8745** with **X-API-Key** authentication.\n* Private, internal model servers (1.5B and Large).\n* Prebaked tags embed model weights for faster startup.\n* Built‑in **CORS** for browser‑based apps.\n* Optional **HTTPS** via built‑in Caddy with ACME.\n\n---\n\n## System Requirements\n\n* **GPU:** NVIDIA with drivers + NVIDIA Container Toolkit.\n* **Base:** CUDA 12.4 (NGC PyTorch).\n* **Driver:** 550+ recommended.\n* **Architectures:** Turing, Ampere, Ada (RTX 20/30/40, A10, L4). Volta (V100) may work.\n* **VRAM (approx.):**\n\n * 1.5B → 10–12 GB.\n * Large → 20–24 GB.\n * Both models enabled concurrently → ~30–36 GB (sum of both, plus minor runtime overhead). On GPUs with ≤24 GB VRAM, prefer running a single model.\n* **Disk (cache):**\n\n * 1.5B → 10–20 GB.\n * Large → 25–40 GB.\n* **Docker:** 24+.\n\nQuick GPU check:\n\n```bash\nnvidia-smi\ndocker run --rm --gpus all nvidia/cuda:12.4.1-base-ubuntu24.04 nvidia-smi\n```\n\n---\n\n## Installation\n\n### Pull\n\n```bash\ndocker pull eworkerinc/vibevoice:latest # alias of :full (prebaked)\n```\n\n### Quick Start (Single Port 8745)\n\nPersistent cache/state setup:\n\n```bash\nsudo mkdir -p /mnt/vv-hf /mnt/vv-voices /mnt/vv-state \\\n && sudo chown -R \"$USER\":\"$USER\" /mnt/vv-hf /mnt/vv-voices /mnt/vv-state\n```\n\nRun both models:\n\n```bash\ndocker run -d --name vibevoice --gpus all \\\n -p 8745:8745 \\\n -v /mnt/vv-hf:/root/.cache/huggingface \\\n -e ENABLE_1_5B=true -e ENABLE_LARGE=true \\\n -e AUTH_REQUIRED=true \\\n -e CORS_ENABLED=true -e ALLOWED_ORIGINS='*' \\\n -v /mnt/vv-voices:/app/voices \\\n -v /mnt/vv-state:/var/lib/eworker \\\n eworkerinc/vibevoice:latest\n```\n\nVRAM note:\n- Enabling both models loads both into GPU memory. Expect ~30–36 GB combined. If you have a 16–24 GB card, use a single model instead (see below).\n\nRun with 1.5B only:\n\n```bash\ndocker run -d --name vibevoice-1_5b --gpus all \\\n -p 8745:8745 -v /mnt/vv-hf:/root/.cache/huggingface \\\n -e ENABLE_1_5B=true -e ENABLE_LARGE=false \\\n eworkerinc/vibevoice:latest\n```\n\nVRAM savings:\n- 1.5B typically fits in ~10–12 GB VRAM, saving roughly 8–12 GB compared to Large. This is the recommended mode for 12–16 GB consumer GPUs and for maximizing concurrency.\n\nRun with Large only:\n\n```bash\ndocker run -d --name vibevoice-large --gpus all \\\n -p 8745:8745 -v /mnt/vv-hf:/root/.cache/huggingface \\\n -e ENABLE_1_5B=false -e ENABLE_LARGE=true \\\n eworkerinc/vibevoice:latest\n```\n\nVRAM trade‑off:\n- Large delivers higher quality at the cost of VRAM (~20–24 GB). Use this mode on 24 GB+ GPUs. Compared to running both, Large‑only reduces memory by ~10–12 GB; compared to 1.5B, it consumes ~8–12 GB more.\n\n### HTTPS with ACME (Optional)\n\n```bash\ndocker run -d --name vibevoice \\\n -p 80:80 -p 443:443 \\\n -v /mnt/vv-hf:/root/.cache/huggingface \\\n -v /mnt/vv-state:/var/lib/eworker \\\n -e TLS_MODE=acme -e TLS_DOMAIN=voice.example.com -e TLS_EMAIL=admin@example.com \\\n -e AUTH_REQUIRED=true -e ENABLE_1_5B=true -e ENABLE_LARGE=true \\\n eworkerinc/vibevoice:latest\n```\n\n\n---\n\n## Authentication & Security\n\n* **Header:** `X-API-Key: ` (printed on first run, persisted under `/var/lib/eworker/api.key`).\n* Rotate keys: `RESET_API_KEY=true` once or set `API_KEY` explicitly.\n* **CORS:** `CORS_ENABLED` + `ALLOWED_ORIGINS`.\n* **HTTPS:** enable ACME with `TLS_MODE=acme`.\n\n---\n\n## Endpoints (Proxy)\n\n* `GET /health` → health probe.\n* `GET /v1/voice/models` → list available models.\n* `GET /v1/voice/models/{model_id}` → model defaults/hints.\n* `GET /v1/voice/voices` → list voices (supports `?model=` filter).\n* `GET /v1/voice/voices/{voice_id}/preview?text=...` → quick WAV preview.\n* `POST /v1/voice/jobs` → start TTS job (returns job\\_id).\n* `GET /v1/voice/jobs` → list jobs; filter by model/status.\n* `GET /v1/voice/jobs/{job_id}` → job status/progress.\n* `GET /v1/voice/jobs/{job_id}/result` → audio result.\n* `POST /v1/voice/jobs/{job_id}/cancel` → cancel job.\n* `GET /v1/voice/jobs/metrics` → aggregated job metrics.\n\n---\n\n## Samples\n\nHealth check:\n\n```bash\ncurl -s http://localhost:8745/health | jq .\n```\n\nGet API key:\n\n```bash\nKEY=$(docker logs vibevoice 2>&1 | sed -n 's/^X-API-Key: //p' | tail -1)\n```\n\nList voices:\n\n```bash\ncurl -s http://localhost:8745/v1/voice/voices -H \"X-API-Key: $KEY\" | jq\n# If only Large is enabled, include the model filter:\ncurl -s \"http://localhost:8745/v1/voice/voices?model=vibevoice-large\" -H \"X-API-Key: $KEY\" | jq\n```\n\nPreview a single voice:\n\n```bash\ncurl -s \"http://localhost:8745/v1/voice/voices/Alice/preview?text=Hello\" \\\n -H \"X-API-Key: $KEY\" --output preview.wav\n```\n\nTwo‑speaker dialogue (very small):\n\n```bash\ncat > body.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: Hello there!\\nSpeaker 2: Hi! Great to meet you.\",\n \"speakers\": [ { \"voiceName\": \"Alice\" }, { \"voiceName\": \"Carter\" } ],\n \"overrides\": {\n \"guidance\": { \"inference_steps\": 28, \"cfg_scale\": 4.5 }\n }\n}\nJSON\n\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @body.json | jq -r .job_id)\n\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > out.wav\n```\n\nCustom voice from external file (mounted WAV):\n\n```bash\n# 1) Place your consented voice sample WAV on the host (3–10 seconds is ideal)\n# Requirements: mono PCM, 16‑bit, 16k or 24k sample rate recommended.\n\n# 2) Copy or move it into the mounted voices folder (host side)\ncp ~/Downloads/my-voice.wav /mnt/vv-voices/\n\n# 3) List voices (the filename without .wav becomes the voice name)\ncurl -s http://localhost:8745/v1/voice/voices -H \"X-API-Key: $KEY\" | jq\n\n# 4) Preview the custom voice (replace MyVoice with your filename stem)\ncurl -s \"http://localhost:8745/v1/voice/voices/MyVoice/preview?text=Hello%20from%20my%20custom%20voice\" \\\n -H \"X-API-Key: $KEY\" --output custom-preview.wav\n\n# 5) Use it in a job\ncat > job-custom.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: This is my custom narrator.\",\n \"speakers\": [ { \"voiceName\": \"MyVoice\" } ],\n \"overrides\": { \"guidance\": { \"inference_steps\": 24, \"cfg_scale\": 4.2 } }\n}\nJSON\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @job-custom.json | jq -r .job_id)\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > custom-out.wav\n```\n\nTiny single‑speaker sample:\n\n```bash\ncat > tiny.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: Hello from VibeVoice.\",\n \"speakers\": [ { \"voiceName\": \"Alice\" } ],\n \"overrides\": { \"guidance\": { \"inference_steps\": 20, \"cfg_scale\": 4.0 } }\n}\nJSON\n\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @tiny.json | jq -r .job_id)\n\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > tiny.wav\n```\n\nTiny two‑speaker sample:\n\n```bash\ncat > tiny2.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: Good morning!\\nSpeaker 2: Morning—ready to start?\",\n \"speakers\": [ { \"voiceName\": \"Mary\" }, { \"voiceName\": \"Frank\" } ],\n \"overrides\": { \"guidance\": { \"inference_steps\": 22, \"cfg_scale\": 4.2 } }\n}\nJSON\n\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @tiny2.json | jq -r .job_id)\n\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > tiny2.wav\n```\n\nFour‑speaker English (~1–2 minutes):\n\n```bash\ncat > roundtable_en.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: Welcome everyone to our quick roundtable.\\nSpeaker 2: Thanks—let's share highlights from this week.\\nSpeaker 3: I experimented with new prompt flows and got cleaner outputs.\\nSpeaker 4: Nice. I focused on latency and shaved a few seconds off.\\nSpeaker 1: Did the new sampler help with stability?\\nSpeaker 2: Yes, especially around tricky words and punctuation.\\nSpeaker 3: I also tried a different CFG scale for warmer tone.\\nSpeaker 4: And fewer inference steps were fine for short clips.\\nSpeaker 1: What about longer narrations?\\nSpeaker 2: We should raise steps slightly to keep quality.\\nSpeaker 3: Agreed—thirty two felt like a sweet spot.\\nSpeaker 4: We could A/B test with audience feedback.\\nSpeaker 1: Let's line up a small user study next week.\\nSpeaker 2: I can prepare scripts and scenarios.\\nSpeaker 3: I will track timing and perceived naturalness.\\nSpeaker 4: I'll measure CPU and GPU utilization.\\nSpeaker 1: Great. Any blockers we should address?\\nSpeaker 2: None on my side.\\nSpeaker 3: All good here.\\nSpeaker 4: Same—excited to ship improvements.\\nSpeaker 1: Perfect—thanks everyone for the quick sync.\",\n \"speakers\": [\n { \"voiceName\": \"Alice\" },\n { \"voiceName\": \"Carter\" },\n { \"voiceName\": \"Frank\" },\n { \"voiceName\": \"Mary\" }\n ],\n \"overrides\": { \"guidance\": { \"inference_steps\": 28, \"cfg_scale\": 4.5 } }\n}\nJSON\n\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @roundtable_en.json | jq -r .job_id)\n\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > roundtable_en.wav\n```\n\nEnglish + Chinese (~1–2 minutes):\n\n```bash\ncat > bilingual.json <<'JSON'\n{\n \"model\": \"vibevoice-1.5b\",\n \"script\": \"Speaker 1: Hello and welcome to our bilingual demo.\\nSpeaker 2: 大家好，欢迎来到我们的双语演示。\\nSpeaker 1: We'll alternate between English and Chinese to compare styles.\\nSpeaker 2: 我会用较自然的语气来朗读示例句子。\\nSpeaker 1: First, a short introduction to the topic.\\nSpeaker 2: 接下来，我们会用简短的段落来说明要点。\\nSpeaker 1: Pay attention to clarity and pacing across languages.\\nSpeaker 2: 请留意语速与停顿是否听起来自然。\\nSpeaker 1: Finally, we'll wrap up with a brief summary.\\nSpeaker 2: 最后，我们会做一个简短的总结。\\nSpeaker 1: Thanks for listening.\\nSpeaker 2: 谢谢收听。\",\n \"speakers\": [ { \"voiceName\": \"Alice\" }, { \"voiceName\": \"Xinran\" } ],\n \"overrides\": { \"guidance\": { \"inference_steps\": 28, \"cfg_scale\": 4.5 } }\n}\nJSON\n\nJOB_ID=$(curl -s -X POST http://localhost:8745/v1/voice/jobs \\\n -H \"Content-Type: application/json\" -H \"X-API-Key: $KEY\" \\\n --data-binary @bilingual.json | jq -r .job_id)\n\ncurl -s \"http://localhost:8745/v1/voice/jobs/$JOB_ID/result\" -H \"X-API-Key: $KEY\" \\\n | jq -r .audio_wav_base64 | base64 --decode > bilingual.wav\n```\n\nNote on Large model:\n- To render with the Large model, set `\"model\": \"vibevoice-large\"` and consider `\"inference_steps\": 40` for best quality.\n\n---\n\n## Built‑in Voices\n\nThe image includes demo voices under `/app/voices` for quick testing. Friendly names:\n\n* Alice (en, woman)\n* Carter (en, man)\n* Frank (en, man)\n* Mary (en, woman, bgm)\n* Maya (en, woman)\n* Samuel (id, man)\n* Anchen (zh, man, bgm)\n* Bowen (zh, man)\n* Xinran (zh, woman)\n\nUse any of the above in `voiceName` or with the preview endpoint (e.g., `/v1/voice/voices/Frank/preview`).\n\n---\n\n## Custom Voices & Consent\n\n- You can add your own voice samples by placing `.wav` files into the mounted voices directory on the host. The container scans `/app/voices` and exposes each file by its filename (without extension) as a `voiceName`.\n- Host mount: `/mnt/vv-voices` → container path `/app/voices` (see run commands above).\n- Recommended sample: 3–10 seconds, mono PCM 16‑bit at 16 kHz or 24 kHz. Trim silence for best results.\n- Use `GET /v1/voice/voices` to confirm your custom voice name, then preview via `GET /v1/voice/voices/{voiceName}/preview` and reference it in jobs (`speakers[].voiceName`).\n- Consent: Only use voice samples you have the legal right and consent to use. If the voice belongs to a third party, obtain explicit permission. You are responsible for complying with applicable laws, terms, and policies.\n\nNote: Advanced users may directly supply absolute `voice_sample_paths` to the upstream API inside the container. The public proxy focuses on folder‑based voice management via `/app/voices` for simplicity and safety.\n\n---\n\n## Environment Variables\n\n* **Auth:** `AUTH_REQUIRED`, `API_KEY`, `RESET_API_KEY`\n* **Models:** `ENABLE_1_5B`, `ENABLE_LARGE`\n* **CORS:** `CORS_ENABLED`, `ALLOWED_ORIGINS`\n* **HF:** `HF_HOME`, `HF_HUB_ENABLE_HF_TRANSFER`, `HUGGING_FACE_HUB_TOKEN`\n* **Attention impl:** `VIBEVOICE_ATTN_IMPL` (default `sdpa` per image env; app falls back to `flash_attention_2` only if the env var is unset)\n* **TLS:** `TLS_MODE`, `TLS_DOMAIN`, `TLS_EMAIL`\n\n---\n\n## Volumes\n\n* `/root/.cache/huggingface` → model cache (mount to persist downloads)\n* `/var/lib/eworker` → state (API key, job state)\n* `/app/voices` → voices folder (.wav files)\n\n---\n\n## Prebaked Weights\n\nThe unified image optionally bakes model weights into the Hugging Face cache inside the image for faster cold starts.\n\n- Cache root: `/root/.cache/huggingface/hub`\n- Models cached:\n - `/root/.cache/huggingface/hub/models--microsoft--VibeVoice-1.5B`\n - `/root/.cache/huggingface/hub/models--microsoft--VibeVoice-Large`\n\nVerify inside the image:\n\n```bash\ndocker run --rm eworkerinc/vibevoice:latest bash -lc \\\n 'ls -1 /root/.cache/huggingface/hub | grep models--microsoft--VibeVoice'\n```\n\nNote: Baked weights reduce first-run downloads but do not change VRAM requirements; enabling both models still sums VRAM usage at runtime (~30–36 GB combined).\n\n---\n\n## Tuning Tips\n\n* `inference_steps` (20–40 typical): higher improves quality.\n* `cfg_scale` (3.5–6.0 typical): adjust style/steadiness.\n* `sample_rate_hz`: 16000 or 24000.\n* Use 1.5B model if VRAM is constrained (saves ~8–12 GB vs Large).\n\n---\n\n## Troubleshooting\n\n* **GPU kernels:** Default SDPA avoids kernel/driver mismatch issues.\n* **VRAM:** If out of memory, disable Large or reduce `inference_steps`/`cfg_scale`.\n* **Disk:** Mount HF cache on large disk; prune Docker caches.\n---\n\n## Integration (E‑Worker Soundstage)\n\n* **Base URL:** `http://localhost:8745`\n* **Header:** `X-API-Key: `\n\n","has_starred","hub_user","immutable_tags_settings",{"_76":11,"_77":78},"enabled","rules",[79],".*","is_automated","permissions",{"_83":17,"_84":11,"_85":11},"read","write","admin","user","storage_size",68444734228,"orgData","recentTags",[92,129],{"_93":45,"_94":95,"_96":97,"_98":99,"_100":101,"_102":103,"_51":118,"_119":95,"_120":29,"_121":57,"_39":122,"_123":124,"_125":109,"_126":118,"_127":66,"_128":17},"content_type","creator",36311160,"digest","sha256:13a444c71be71a6c28e03a8d55222434fb240af5a76620887bce722e8c91a347","full_size",37060859399,"id",966542838,"images",[104],{"_105":106,"_96":97,"_107":49,"_108":109,"_110":111,"_112":113,"_114":49,"_115":-5,"_116":99,"_63":66,"_117":-5},"architecture","amd64","features","last_pulled","2026-09-23T10:04:17.973927158Z","last_pushed","2025-09-03T18:04:18.104932804Z","os","linux","os_features","os_version","size","variant","2025-09-03T18:04:18.87795Z","last_updater","last_updater_username","media_type","latest","repository",28455027,"tag_last_pulled","tag_last_pushed","tag_status","v2",{"_93":45,"_94":95,"_96":97,"_98":99,"_100":130,"_102":131,"_51":133,"_119":95,"_120":29,"_121":57,"_39":134,"_123":124,"_125":109,"_126":133,"_127":66,"_128":17},966545274,[132],{"_105":106,"_96":97,"_107":49,"_108":109,"_110":111,"_112":113,"_114":49,"_115":-5,"_116":99,"_63":66,"_117":-5},"2025-09-03T18:08:27.65853Z","full","sbxKitMetadata","isPotentialSpam","isHubSbxKitsEnabled","showSbxLogoOverride","sbxTrademarkText","displaySecurityMetadata","tagAttestations","tagsHaveAttestations",[],"routes/_layout.r.$namespace.$reponame.tags",{"_34":146,"_154":155,"_142":168,"_169":11},{"_36":147,"_43":149,"_46":47,"_48":49,"_50":11,"_51":52,"_53":54,"_55":150,"_39":31,"_28":29,"_58":59,"_60":45,"_61":62,"_63":64,"_65":66,"_67":-5,"_68":69,"_70":71,"_72":11,"_73":29,"_74":151,"_80":11,"_81":153,"_86":29,"_87":88},[148],{"_39":40,"_41":42},[45],[57],{"_76":11,"_77":152},[79],{"_83":17,"_84":11,"_85":11},"tags",{"_156":157,"_158":-5,"_159":-5,"_160":161},"count",2,"next","previous","results",[162,165],{"_93":45,"_94":95,"_96":97,"_98":99,"_100":130,"_102":163,"_51":133,"_119":95,"_120":29,"_121":57,"_39":134,"_123":124,"_125":109,"_126":133,"_127":66,"_128":17},[164],{"_105":106,"_96":97,"_107":49,"_108":109,"_110":111,"_112":113,"_114":49,"_115":-5,"_116":99,"_63":66,"_117":-5},{"_93":45,"_94":95,"_96":97,"_98":99,"_100":101,"_102":166,"_51":118,"_119":95,"_120":29,"_121":57,"_39":122,"_123":124,"_125":109,"_126":118,"_127":66,"_128":17},[167],{"_105":106,"_96":97,"_107":49,"_108":109,"_110":111,"_112":113,"_114":49,"_115":-5,"_116":99,"_63":66,"_117":-5},[],"canDeleteTags","actionData","errors"]

eworkerinc/vibevoice

Tags for eworkerinc/vibevoice