{"id":2570,"date":"2026-06-14T16:39:56","date_gmt":"2026-06-14T16:39:56","guid":{"rendered":"https:\/\/ai-box.eu\/?p=2570"},"modified":"2026-06-14T16:50:04","modified_gmt":"2026-06-14T16:50:04","slug":"nvidia-magpie-tts-locally-german-speech-output-as-a-microservice","status":"publish","type":"post","link":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/","title":{"rendered":"NVIDIA Magpie TTS locally: German speech output as a microservice"},"content":{"rendered":"<p>The speech recognition is in place: with Parakeet (<a href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-nim-locally-running-german-speech-recognition-as-a-microservice\/2556\/\" target=\"_blank\" rel=\"noopener\">Part 2<\/a>) and Canary (<a href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-canary-locally-multilingual-speech-recognition-and-translation-as-a-nim\/2562\/\" target=\"_blank\" rel=\"noopener\">Part 3<\/a>) I covered the input direction, speech to text, natively on NVIDIA. Now comes the opposite direction, the <strong>voice output<\/strong>. In this post I run <strong>NVIDIA Magpie TTS<\/strong> as a local NIM and have German text read aloud naturally. This is the NVIDIA-native counterpart to my earlier post on <a href=\"https:\/\/ai-box.eu\/ai-pipeline\/open-webui-sprachsteuerung-lokal-deutsche-stt-tts-mit-whisper-piper-und-xtts\/2490\/\" target=\"_blank\" rel=\"noopener\">German TTS with Piper and XTTS<\/a>, but this time entirely within the NVIDIA ecosystem and, as always, local, on my own hardware.<\/p>\n<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_85 counter-hierarchy ez-toc-counter ez-toc-grey ez-toc-container-direction\">\n<div class=\"ez-toc-title-container\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">Table of Contents<\/p>\n<span class=\"ez-toc-title-toggle\"><a href=\"#\" class=\"ez-toc-pull-right ez-toc-btn ez-toc-btn-xs ez-toc-btn-default ez-toc-toggle\" aria-label=\"Toggle Table of Content\"><span class=\"ez-toc-js-icon-con\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #999;color:#999\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #999;color:#999\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/span><\/a><\/span><\/div>\n<nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#What_is_Magpie_TTS\" >What is Magpie TTS?<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#The_goal_of_this_post\" >The goal of this post<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Requirements\" >Requirements<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-4\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_1_Start_the_Magpie_TTS_NIM\" >Step 1: Start the Magpie TTS NIM<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-5\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_2_Check_the_container_status\" >Step 2: Check the container status<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-6\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_3_List_the_available_voices\" >Step 3: List the available voices<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-7\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_4_First_German_speech_output\" >Step 4: First German speech output<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-8\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_5_Streaming_TTS_for_low_latency\" >Step 5: Streaming TTS for low latency<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-9\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Step_6_Voices_and_styles\" >Step 6: Voices and styles<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-10\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Tips_and_troubleshooting\" >Tips and troubleshooting<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-11\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#Conclusion\" >Conclusion<\/a><\/li><\/ul><\/nav><\/div>\n<h2><span class=\"ez-toc-section\" id=\"What_is_Magpie_TTS\"><\/span>What is Magpie TTS?<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Magpie TTS is an <strong>end-to-end multilingual neural text-to-speech model<\/strong>. It generates speech by predicting discrete audio codec tokens via a transformer encoder-decoder architecture; a downstream audio codec model then turns those tokens into the audible waveform.<\/p>\n<p>Three properties matter for us:<\/p>\n<ul>\n<li><strong>Multilingual incl. German:<\/strong> Magpie Multilingual covers nine languages, German (<code>de-DE<\/code>) among them.<\/li>\n<li><strong>Streaming and offline:<\/strong> it can deliver the finished audio in one piece or stream the first fragments as soon as they are ready. The latter matters a lot to me for the later voice-agent feel.<\/li>\n<li><strong>Multiple voices:<\/strong> at least one male and one female voice per language, partly with emotional styles.<\/li>\n<\/ul>\n<p>In short: the TTS counterpart to the ASR NIMs from Parts 2 and 3.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"The_goal_of_this_post\"><\/span>The goal of this post<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>We run the Magpie TTS NIM locally: <strong>German text in, natural German speech as a WAV out<\/strong>. Optionally also in streaming mode, which delivers the first audio fragments as soon as they are ready. Everything local, so your text and the generated audio stay on the machine.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Requirements\"><\/span>Requirements<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>If you have been through Parts 2 and 3, the groundwork is already in place and we only reference it briefly:<\/p>\n<ul>\n<li>NGC account and API key, <code>docker login<\/code> to <code>nvcr.io<\/code> (see Part 2)<\/li>\n<li>the <code>riva-client<\/code> venv with <code>nvidia-riva-client<\/code> installed \u2013 the cloned <code>python-clients<\/code> repo also includes the TTS scripts under <code>scripts\/tts\/<\/code><\/li>\n<li>GPU \u2265 compute capability 8.0 \u2013 Magpie Multilingual uses about 11 GB VRAM at <code>batch_size=8<\/code><\/li>\n<\/ul>\n<p>So you don&#8217;t need to install anything new. <strong>One important point:<\/strong> Magpie uses the same ports (9000\/50051) as the ASR NIMs. Before this step, stop any Parakeet or Canary container that may still be running (<code>Ctrl + C<\/code>), otherwise you get a port conflict.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_1_Start_the_Magpie_TTS_NIM\"><\/span>Step 1: Start the Magpie TTS NIM<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>If your API key is no longer set in the current terminal session, set it again:<\/p>\n<p><strong>Command:<\/strong> <code>export NGC_API_KEY=\"nvapi-xxxxxxxxxxxxxxxxxxxxx\"<\/code><\/p>\n<p>Then pick the container and profile. For the TTS NIM the profile is selected via <code>name=<\/code> (not via <code>mode=<\/code> like for ASR):<\/p>\n<p><strong>Command:<\/strong> <code>export CONTAINER_ID=magpie-tts-multilingual<\/code><\/p>\n<p><strong>Command:<\/strong> <code>export NIM_TAGS_SELECTOR=name=magpie-tts-multilingual<\/code><\/p>\n<p><strong>Command:<\/strong> <code>docker run -it --rm --name=$CONTAINER_ID --runtime=nvidia --gpus '\"device=0\"' --shm-size=8GB -e NGC_API_KEY -e NIM_HTTP_API_PORT=9000 -e NIM_GRPC_API_PORT=50051 -p 9000:9000 -p 50051:50051 -e NIM_TAGS_SELECTOR -v ~\/.cache\/nim:\/opt\/nim\/.cache nvcr.io\/nim\/nvidia\/$CONTAINER_ID:latest<\/code><\/p>\n<p>You can reuse the cache directory <code>~\/.cache\/nim<\/code> from Parts 2\/3. The first start again downloads the model and builds the inference engine; the service is ready once &#8220;Application is ready to receive API requests&#8221; appears in the logs.<\/p>\n<p><strong>Note on batch size:<\/strong><\/p>\n<p>The default is <code>batch_size=8<\/code> (~11 GB VRAM). If you need more parallelism, append the batch size: <code>export NIM_TAGS_SELECTOR=\"name=magpie-tts-multilingual,batch_size=32\"<\/code>. That configuration uses about 31 GB, though. On the A6000 Ada (48 GB) both are feasible; keep an eye on usage with <code>nvidia-smi<\/code>.<\/p>\n<p>After the start, the terminal again looked rather unspectacular. But everything should be fine and properly up.<\/p>\n<div id=\"attachment_2564\" style=\"width: 999px\" class=\"wp-caption alignnone\"><a href=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg\"><img loading=\"lazy\" decoding=\"async\" aria-describedby=\"caption-attachment-2564\" class=\"size-full wp-image-2564\" src=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg\" alt=\"NVIDIA NIM container Magpie setup\" width=\"989\" height=\"530\" srcset=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg 989w, https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07-300x161.jpg 300w, https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07-768x412.jpg 768w\" sizes=\"(max-width: 989px) 100vw, 989px\" \/><\/a><p id=\"caption-attachment-2564\" class=\"wp-caption-text\">NVIDIA NIM container Magpie setup<\/p><\/div>\n<h2><span class=\"ez-toc-section\" id=\"Step_2_Check_the_container_status\"><\/span>Step 2: Check the container status<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>In a second terminal you check as usual whether the service is running.<\/p>\n<p><strong>Command:<\/strong> <code>docker ps<\/code><\/p>\n<p><strong>Command:<\/strong> <code>curl http:\/\/localhost:9000\/v1\/health\/ready<\/code><\/p>\n<p>If the health check answers with <code>{\"object\":\"health.response\",\"message\":\"ready\",\"status\":\"ready\"}<\/code>, your TTS microservice is up.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Step_3_List_the_available_voices\"><\/span>Step 3: List the available voices<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Before we have anything read aloud, we look at which voices the model offers. First activate the venv:<\/p>\n<p><strong>Command:<\/strong> <code>source ~\/venvs\/riva-client\/bin\/activate<\/code><\/p>\n<p>Then query the voices:<\/p>\n<p><strong>Command:<\/strong> <code>python python-clients\/scripts\/tts\/talk.py --server 0.0.0.0:50051 --list-voices<\/code><\/p>\n<p>You get back a JSON list of languages and voices. For German, look for entries that start with <code>Magpie-Multilingual.DE-DE.<\/code>. There are several voices (for example a female and a male one), partly with emotional styles (e.g. <code>.Neutral<\/code>, <code>.Calm<\/code>). We enter the exact name of the German voice you want into the synthesis command in a moment.<\/p>\n<p>Here is an excerpt of the voices for the German language:<\/p>\n<blockquote><p>&#8220;Magpie-Multilingual.IT-IT.Pascal.Happy&#8221;,<br \/>\n&#8220;Magpie-Multilingual.IT-IT.Pascal.Disgust&#8221;,<br \/>\n&#8220;Magpie-Multilingual.IT-IT.Pascal.Sad&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Neutral&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Calm&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Angry&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Happy&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Disgust&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Pascal.Sad&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia.Neutral&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia.Calm&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia.Angry&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia.Happy&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Mia.Sad&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.Neutral&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.Calm&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.Angry&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.Happy&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.PleasantSurprised&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Diego.Disgust&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia.Neutral&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia.Calm&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia.Angry&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia.Happy&#8221;,<br \/>\n&#8220;Magpie-Multilingual.DE-DE.Sofia.Fearful&#8221;,<br \/>\n&#8220;Magpie-Multilingual.EN-US.Pascal&#8221;,<br \/>\n&#8220;Magpie-Multilingual.EN-US.Pascal.Neutral&#8221;,<br \/>\n&#8220;Magpie-Multilingual.EN-US.Pascal.Calm&#8221;,<br \/>\n&#8220;Magpie-Multilingual.EN-US.Pascal.Angry&#8221;,<\/p><\/blockquote>\n<h2><span class=\"ez-toc-section\" id=\"Step_4_First_German_speech_output\"><\/span>Step 4: First German speech output<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Now we have German text read aloud and save the result as a WAV. Important on a <strong>headless server<\/strong> (like my A6000 Ada over SSH): there is no sound card for direct playback, and I&#8217;m not sitting at the machine anyway. So we write to a file (<code>--output<\/code>) and listen to it afterwards (e.g. after downloading via <code>scp<\/code>).<\/p>\n<p><strong>Note:<\/strong> Please make sure you adjust the <code>--output<\/code> parameter so the file is written to a path that actually exists on your system.<\/p>\n<p><strong>Command:<\/strong> <code>python python-clients\/scripts\/tts\/talk.py --server 0.0.0.0:50051 --language-code de-DE --voice Magpie-Multilingual.DE-DE.Sofia.Neutral --text \"Hallo, das ist eine lokale Sprachausgabe mit NVIDIA Magpie.\" --output \/home\/ingmar\/asr\/de_Sofia.Neutral.wav<\/code><\/p>\n<p>You replace the voice name with the appropriate value from Step 3.<\/p>\n<p>Alternatively, this also works directly via the NIM&#8217;s HTTP interface:<\/p>\n<p><strong>Command:<\/strong> <code>curl -sS http:\/\/localhost:9000\/v1\/audio\/synthesize --fail-with-body -F language=de-DE -F text=\"Hallo, das ist eine lokale Sprachausgabe mit NVIDIA Magpie.\" --output ausgabe.wav<\/code><\/p>\n<p><strong>Audio sample:<\/strong><\/p>\n<audio class=\"wp-audio-shortcode\" id=\"audio-2570-1\" preload=\"none\" style=\"width: 100%;\" controls=\"controls\"><source type=\"audio\/wav\" src=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/de_PleasantSurprised.wav?_=1\" \/><a href=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/de_PleasantSurprised.wav\">https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/de_PleasantSurprised.wav<\/a><\/audio>\n<h2><span class=\"ez-toc-section\" id=\"Step_5_Streaming_TTS_for_low_latency\"><\/span>Step 5: Streaming TTS for low latency<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>A note up front, because &#8220;streaming&#8221; is slightly confusing here: no URL is opened in the browser and nothing is played &#8220;live&#8221;. It is exactly the same command-line call as in Step 4, just with the additional <code>--stream<\/code> switch. The difference lies solely in <em>how<\/em> the NIM returns the audio:<\/p>\n<ul>\n<li><strong>Without <code>--stream<\/code> (offline):<\/strong> you send the text, the NIM synthesizes the <strong>complete<\/strong> speech and only then returns it in one piece. You wait until everything is done.<\/li>\n<li><strong>With <code>--stream<\/code>:<\/strong> the NIM delivers the audio in <strong>small fragments<\/strong> as soon as they are ready. The client already gets the first fractions of a second while the rest is still being generated. That is the low &#8220;time-to-first-audio&#8221; that makes a voice agent feel natural.<\/li>\n<\/ul>\n<p>So for now it is only about testing whether the <code>--stream<\/code> parameter works at all.<\/p>\n<p><strong>Command:<\/strong> <code>python python-clients\/scripts\/tts\/talk.py --server 0.0.0.0:50051 --language-code de-DE --voice Magpie-Multilingual.DE-DE.Mia.Calm --text \"Hallo, das ist eine lokale Sprachausgabe mit NVIDIA Magpie.\" --stream --output \/home\/ingmar\/asr\/ausgabe_stream.wav<\/code><\/p>\n<p><strong>What does that mean on a headless server in practice?<\/strong><\/p>\n<p>In both cases the same file <code>ausgabe*.wav<\/code> is produced at the end, which you download and listen to. You only <em>hear<\/em> the streaming advantage (start talking immediately instead of waiting) when the audio goes to a speaker live. Here on the server we mainly test <strong>that<\/strong> the streaming mode runs cleanly and produces a valid WAV. Streaming shows its real benefit later in Part 6, when Pipecat passes the audio fragments straight to playback.<\/p>\n<p>Two practical points:<\/p>\n<ul>\n<li><strong>Listening:<\/strong> download the WAV via <code>scp<\/code> to your workstation and play it there.<\/li>\n<li><strong>Long texts:<\/strong> for long inputs, <code>--stream<\/code> is the better choice anyway, because otherwise the offline response can exceed the gRPC limit of 4 MB per message.<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"Step_6_Voices_and_styles\"><\/span>Step 6: Voices and styles<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>Magpie Multilingual brings at least one male and one female voice per language, plus emotional styles. Via the <code>--voice<\/code> name you select exactly which voice and which style is spoken. We have already seen that.<\/p>\n<p><strong>Note on voice cloning:<\/strong> True zero-shot voice cloning (reproducing a voice from a short audio prompt) is <em>not<\/em> part of this multilingual NIM, but a separate, access-restricted model (<code>magpie-tts-zeroshot<\/code>). Here we use the predefined voices. They are perfectly sufficient for testing and experimenting with a clean German voice agent. Whether you really want to use them in production is up to you.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Tips_and_troubleshooting\"><\/span>Tips and troubleshooting<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<ul>\n<li><strong>Port conflict:<\/strong> Magpie occupies 9000\/50051 like the ASR NIMs. Stop the other container first, or assign different ports in the <code>docker run<\/code> command.<\/li>\n<li><strong>Headless server:<\/strong> No direct playback without a sound card \u2013 write the audio output to a WAV with <code>--output<\/code> and listen to it after downloading (instead of <code>--play-audio<\/code>).<\/li>\n<li><strong>Long texts \/ gRPC 4 MB limit:<\/strong> For long inputs use the streaming mode (<code>--stream<\/code>), otherwise the offline response can blow past the gRPC message size.<\/li>\n<li><strong>Copy the voice name exactly:<\/strong> Copy the <code>--voice<\/code> value exactly from <code>--list-voices<\/code>; a typo leads to an error.<\/li>\n<li><strong>VRAM:<\/strong> ~11 GB at <code>batch_size=8<\/code>, ~31 GB at <code>batch_size=32<\/code>. Check with <code>nvidia-smi<\/code>.<\/li>\n<li><strong>Sample rate\/encoding:<\/strong> The default is <code>LINEAR_PCM<\/code>; via the parameters you can control, for example, the output sample rate (around 44100 Hz).<\/li>\n<\/ul>\n<h2><span class=\"ez-toc-section\" id=\"Conclusion\"><\/span>Conclusion<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>With Magpie, the <strong>voice output<\/strong> is now available as an NVIDIA microservice too. That gives me <strong>ASR<\/strong> (Parakeet, Canary) and <strong>TTS<\/strong> (Magpie) together, fully NVIDIA-native and local. These are the two halves of a voice agent: listening and answering.<\/p>\n<p>In the next part comes the <strong>brain<\/strong>: the orchestrator via the <strong>NVIDIA NeMo Agent Toolkit (NAT)<\/strong>, which recognizes what the user wants and triggers the matching action. After that, in Part 6, we connect all the building blocks with Pipecat into a continuous, interruptible voice loop \u2013 the complete local voice agent.<\/p>\n<p>If you rebuild the setup: drop me a comment about which German Magpie voice sounds most natural to you.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>The speech recognition is in place: with Parakeet (Part 2) and Canary (Part 3) I covered the input direction, speech to text, natively on NVIDIA. Now comes the opposite direction, the voice output. In this post I run NVIDIA Magpie TTS as a local NIM and have German text read aloud naturally. This is the [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":2565,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"_et_pb_use_builder":"","_et_pb_old_content":"","_et_gb_content_width":"","footnotes":""},"categories":[162,8,50],"tags":[353,1658,1657,1031,1418,1662,1665,1608,1623,1625,1620,1661,1656,1617,1621,315,1032,1659,1664,1663,1660,1624],"class_list":["post-2570","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-large-language-models-en","category-news","category-top-story-en","tag-docker","tag-german-speech-synthesis","tag-german-text-to-speech","tag-local-ai","tag-lokale-ki","tag-magpie-tts-nim","tag-magpie-tts-multilingual","tag-ngc-api-key","tag-nvcr-io","tag-nvidia-inference-microservices","tag-nvidia-magpie-tts","tag-nvidia-magpie-tts-locally","tag-nvidia-nim","tag-nvidia-riva","tag-rtx-a6000-en","tag-sovereign-ai","tag-speech-synthesis","tag-streaming-tts","tag-text-to-speech","tag-tts-voices","tag-voice-agent","et-has-post-format-content","et_post_format-et-post-format-standard"],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v27.8 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box<\/title>\n<meta name=\"description\" content=\"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.\" \/>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/\" \/>\n<meta property=\"og:locale\" content=\"en_US\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box\" \/>\n<meta property=\"og:description\" content=\"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.\" \/>\n<meta property=\"og:url\" content=\"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/\" \/>\n<meta property=\"og:site_name\" content=\"Exploring the Future: Inside the AI Box\" \/>\n<meta property=\"article:published_time\" content=\"2026-06-14T16:39:56+00:00\" \/>\n<meta property=\"article:modified_time\" content=\"2026-06-14T16:50:04+00:00\" \/>\n<meta property=\"og:image\" content=\"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg\" \/>\n\t<meta property=\"og:image:width\" content=\"989\" \/>\n\t<meta property=\"og:image:height\" content=\"530\" \/>\n\t<meta property=\"og:image:type\" content=\"image\/jpeg\" \/>\n<meta name=\"author\" content=\"Maker\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:creator\" content=\"@Ingmar_Stapel\" \/>\n<meta name=\"twitter:site\" content=\"@Ingmar_Stapel\" \/>\n<meta name=\"twitter:label1\" content=\"Written by\" \/>\n\t<meta name=\"twitter:data1\" content=\"Maker\" \/>\n\t<meta name=\"twitter:label2\" content=\"Est. reading time\" \/>\n\t<meta name=\"twitter:data2\" content=\"8 minutes\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"Article\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#article\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/\"},\"author\":{\"name\":\"Maker\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/#\\\/schema\\\/person\\\/cc91d08618b3feeef6926591b465eab1\"},\"headline\":\"NVIDIA Magpie TTS locally: German speech output as a microservice\",\"datePublished\":\"2026-06-14T16:39:56+00:00\",\"dateModified\":\"2026-06-14T16:50:04+00:00\",\"mainEntityOfPage\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/\"},\"wordCount\":1460,\"commentCount\":0,\"image\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/ai-box.eu\\\/wp-content\\\/uploads\\\/2026\\\/06\\\/NVIDIA_nim_container_Magpie_setup_07.jpg\",\"keywords\":[\"Docker\",\"German speech synthesis\",\"German text-to-speech\",\"local AI\",\"lokale KI\",\"Magpie TTS NIM\",\"magpie-tts-multilingual\",\"NGC API-Key\",\"NGC API-Key\",\"nvcr.io\",\"NVIDIA Inference Microservices\",\"NVIDIA Magpie TTS\",\"NVIDIA Magpie TTS locally\",\"NVIDIA NIM\",\"NVIDIA Riva\",\"RTX A6000\",\"sovereign AI\",\"speech synthesis\",\"Streaming TTS\",\"Text-to-Speech\",\"TTS voices\",\"Voice Agent\"],\"articleSection\":[\"Large Language Models\",\"News\",\"Top story\"],\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"CommentAction\",\"name\":\"Comment\",\"target\":[\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#respond\"]}]},{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/\",\"url\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/\",\"name\":\"NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/#website\"},\"primaryImageOfPage\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#primaryimage\"},\"image\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#primaryimage\"},\"thumbnailUrl\":\"https:\\\/\\\/ai-box.eu\\\/wp-content\\\/uploads\\\/2026\\\/06\\\/NVIDIA_nim_container_Magpie_setup_07.jpg\",\"datePublished\":\"2026-06-14T16:39:56+00:00\",\"dateModified\":\"2026-06-14T16:50:04+00:00\",\"author\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/#\\\/schema\\\/person\\\/cc91d08618b3feeef6926591b465eab1\"},\"description\":\"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#breadcrumb\"},\"inLanguage\":\"en-US\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/\"]}]},{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#primaryimage\",\"url\":\"https:\\\/\\\/ai-box.eu\\\/wp-content\\\/uploads\\\/2026\\\/06\\\/NVIDIA_nim_container_Magpie_setup_07.jpg\",\"contentUrl\":\"https:\\\/\\\/ai-box.eu\\\/wp-content\\\/uploads\\\/2026\\\/06\\\/NVIDIA_nim_container_Magpie_setup_07.jpg\",\"width\":989,\"height\":530,\"caption\":\"NVIDIA NIM Container Magpie setup\"},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/news\\\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\\\/2570\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Start\",\"item\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"NVIDIA Magpie TTS locally: German speech output as a microservice\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/#website\",\"url\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/\",\"name\":\"Exploring the Future: Inside the AI Box\",\"description\":\"Inside the AI Box, we share our experiences and discoveries in the world of artificial intelligence.\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-US\"},{\"@type\":\"Person\",\"@id\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/#\\\/schema\\\/person\\\/cc91d08618b3feeef6926591b465eab1\",\"name\":\"Maker\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-US\",\"@id\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g\",\"url\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g\",\"contentUrl\":\"https:\\\/\\\/secure.gravatar.com\\\/avatar\\\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g\",\"caption\":\"Maker\"},\"description\":\"I live in Bavaria near Munich. In my head I always have many topics and try out especially in the field of Internet new media much in my spare time. I write on the blog because it makes me fun to report about the things that inspire me. I am happy about every comment, about suggestion and very about questions.\",\"sameAs\":[\"https:\\\/\\\/ai-box.eu\"],\"url\":\"https:\\\/\\\/ai-box.eu\\\/en\\\/author\\\/ingmars\\\/\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box","description":"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/","og_locale":"en_US","og_type":"article","og_title":"NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box","og_description":"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.","og_url":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/","og_site_name":"Exploring the Future: Inside the AI Box","article_published_time":"2026-06-14T16:39:56+00:00","article_modified_time":"2026-06-14T16:50:04+00:00","og_image":[{"width":989,"height":530,"url":"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg","type":"image\/jpeg"}],"author":"Maker","twitter_card":"summary_large_image","twitter_creator":"@Ingmar_Stapel","twitter_site":"@Ingmar_Stapel","twitter_misc":{"Written by":"Maker","Est. reading time":"8 minutes"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"Article","@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#article","isPartOf":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/"},"author":{"name":"Maker","@id":"https:\/\/ai-box.eu\/en\/#\/schema\/person\/cc91d08618b3feeef6926591b465eab1"},"headline":"NVIDIA Magpie TTS locally: German speech output as a microservice","datePublished":"2026-06-14T16:39:56+00:00","dateModified":"2026-06-14T16:50:04+00:00","mainEntityOfPage":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/"},"wordCount":1460,"commentCount":0,"image":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#primaryimage"},"thumbnailUrl":"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg","keywords":["Docker","German speech synthesis","German text-to-speech","local AI","lokale KI","Magpie TTS NIM","magpie-tts-multilingual","NGC API-Key","NGC API-Key","nvcr.io","NVIDIA Inference Microservices","NVIDIA Magpie TTS","NVIDIA Magpie TTS locally","NVIDIA NIM","NVIDIA Riva","RTX A6000","sovereign AI","speech synthesis","Streaming TTS","Text-to-Speech","TTS voices","Voice Agent"],"articleSection":["Large Language Models","News","Top story"],"inLanguage":"en-US","potentialAction":[{"@type":"CommentAction","name":"Comment","target":["https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#respond"]}]},{"@type":"WebPage","@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/","url":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/","name":"NVIDIA Magpie TTS locally: German speech output as a microservice - Exploring the Future: Inside the AI Box","isPartOf":{"@id":"https:\/\/ai-box.eu\/en\/#website"},"primaryImageOfPage":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#primaryimage"},"image":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#primaryimage"},"thumbnailUrl":"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg","datePublished":"2026-06-14T16:39:56+00:00","dateModified":"2026-06-14T16:50:04+00:00","author":{"@id":"https:\/\/ai-box.eu\/en\/#\/schema\/person\/cc91d08618b3feeef6926591b465eab1"},"description":"Run NVIDIA Magpie TTS locally: turn German text into natural speech as a NIM \u2014 multiple voices, streaming-capable and entirely on your own hardware.","breadcrumb":{"@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#breadcrumb"},"inLanguage":"en-US","potentialAction":[{"@type":"ReadAction","target":["https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/"]}]},{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#primaryimage","url":"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg","contentUrl":"https:\/\/ai-box.eu\/wp-content\/uploads\/2026\/06\/NVIDIA_nim_container_Magpie_setup_07.jpg","width":989,"height":530,"caption":"NVIDIA NIM Container Magpie setup"},{"@type":"BreadcrumbList","@id":"https:\/\/ai-box.eu\/en\/news\/nvidia-magpie-tts-locally-german-speech-output-as-a-microservice\/2570\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Start","item":"https:\/\/ai-box.eu\/en\/"},{"@type":"ListItem","position":2,"name":"NVIDIA Magpie TTS locally: German speech output as a microservice"}]},{"@type":"WebSite","@id":"https:\/\/ai-box.eu\/en\/#website","url":"https:\/\/ai-box.eu\/en\/","name":"Exploring the Future: Inside the AI Box","description":"Inside the AI Box, we share our experiences and discoveries in the world of artificial intelligence.","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/ai-box.eu\/en\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-US"},{"@type":"Person","@id":"https:\/\/ai-box.eu\/en\/#\/schema\/person\/cc91d08618b3feeef6926591b465eab1","name":"Maker","image":{"@type":"ImageObject","inLanguage":"en-US","@id":"https:\/\/secure.gravatar.com\/avatar\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g","url":"https:\/\/secure.gravatar.com\/avatar\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g","contentUrl":"https:\/\/secure.gravatar.com\/avatar\/e96b93fc3c7e50c1f21c5c6b1f146dc4867936141360830b328947b32cacf93a?s=96&d=mm&r=g","caption":"Maker"},"description":"I live in Bavaria near Munich. In my head I always have many topics and try out especially in the field of Internet new media much in my spare time. I write on the blog because it makes me fun to report about the things that inspire me. I am happy about every comment, about suggestion and very about questions.","sameAs":["https:\/\/ai-box.eu"],"url":"https:\/\/ai-box.eu\/en\/author\/ingmars\/"}]}},"_links":{"self":[{"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/posts\/2570","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/comments?post=2570"}],"version-history":[{"count":1,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/posts\/2570\/revisions"}],"predecessor-version":[{"id":2571,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/posts\/2570\/revisions\/2571"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/media\/2565"}],"wp:attachment":[{"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/media?parent=2570"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/categories?post=2570"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ai-box.eu\/en\/wp-json\/wp\/v2\/tags?post=2570"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}