voice-ai-workbench 0.2.0__tar.gz

voice_ai_workbench-0.2.0/LICENSE

@@ -0,0 +1,202 @@
+
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright [yyyy] [name of copyright owner]
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

voice_ai_workbench-0.2.0/MANIFEST.in

@@ -0,0 +1,4 @@
+include LICENSE NOTICE
+recursive-include presets *.yaml
+recursive-include src/multi_modal_ai_studio/webui/static *.html *.js *.css
+global-exclude *.backup

voice_ai_workbench-0.2.0/NOTICE

@@ -0,0 +1,7 @@
+Multi-modal AI Studio
+Copyright 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+
+This product includes software developed by Multi-modal AI Studio contributors.
+
+Third-party dependencies and their licenses are documented in
+docs/third_party_dependencies.md.

voice_ai_workbench-0.2.0/PKG-INFO

@@ -0,0 +1,199 @@
+Metadata-Version: 2.4
+Name: voice-ai-workbench
+Version: 0.2.0
+Summary: A developer workbench for voice and optional vision AI, with interchangeable models and conversation latency analysis
+Author-email: "Chitoku YATO (tokk-nv)" <cyato@nvidia.com>, "Aditya Sahu (adsahu-nv)" <adsahu@nvidia.com>, kbenkhaled <kbenkhaled@nvidia.com>
+License: Apache-2.0
+Project-URL: Homepage, https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio
+Project-URL: Documentation, https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/tree/main/docs
+Project-URL: Repository, https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio
+Project-URL: Issues, https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/issues
+Keywords: ai,voice,multimodal,riva,nvidia,openai,conversational-ai
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+License-File: LICENSE
+License-File: NOTICE
+Requires-Dist: aiohttp>=3.14.5
+Requires-Dist: psutil>=5.9.0
+Requires-Dist: grpcio>=1.56.2
+Requires-Dist: nvidia-riva-client>=2.28.0
+Requires-Dist: protobuf>=7.36.2
+Requires-Dist: openai>=1.0.0
+Requires-Dist: pyyaml>=6.0
+Requires-Dist: python-dotenv>=1.2.2
+Requires-Dist: numpy>=1.23.2
+Requires-Dist: websockets>=11.0
+Provides-Extra: dev
+Requires-Dist: pytest>=9.0.3; extra == "dev"
+Requires-Dist: pytest-asyncio>=1.3.0; extra == "dev"
+Requires-Dist: pytest-cov>=7.0.0; extra == "dev"
+Requires-Dist: black>=26.10.1; extra == "dev"
+Requires-Dist: ruff>=0.17.2; extra == "dev"
+Requires-Dist: mypy>=2.4.0; extra == "dev"
+Requires-Dist: uv>=0.13.5; extra == "dev"
+Provides-Extra: audio
+Requires-Dist: pyaudio>=0.2.13; extra == "audio"
+Provides-Extra: camera
+Requires-Dist: opencv-python-headless>=4.8.1.78; extra == "camera"
+Provides-Extra: video-input
+Requires-Dist: av>=10.0.0; extra == "video-input"
+Requires-Dist: Pillow>=13.0.0; extra == "video-input"
+Provides-Extra: webrtc-camera
+Requires-Dist: aiortc>=1.6.0; extra == "webrtc-camera"
+Requires-Dist: av>=10.0.0; extra == "webrtc-camera"
+Requires-Dist: opencv-python-headless>=4.8.1.78; extra == "webrtc-camera"
+Dynamic: license-file
+
+# Voice AI Workbench
+
+**Give your robot a voice. Compare models, shape its character, and measure the conversation.**
+
+![Voice AI Workbench showing a recorded NVIDIA Speech conversation and its latency timeline](https://raw.githubusercontent.com/NVIDIA-AI-IOT/multi_modal_ai_studio/main/docs/images/voice-ai-workbench-nvidia-speech.png)
+
+Recorded on Jetson AGX Thor: Nemotron 3.5 ASR → Gemma 4 E2B → Magpie TTS, using the NVIDIA Speech — Realtime preset. The waveforms, partial transcripts, latency bands, and CPU/GPU traces come from a real session; the internal hostname and IP addresses are omitted from this public screenshot. These timings illustrate one conversation, not a controlled benchmark.
+
+Voice AI Workbench is a developer tool for building and tuning conversational AI. Connect independent speech recognition (ASR), language model (LLM), and text-to-speech (TTS) services; adjust the LLM system prompt; and compare recognition, response latency, playback, and interruptions on a shared timeline. Optional camera input lets you evaluate vision-language models too.
+
+Start with your browser's microphone and speaker for a quick test. Then select a microphone and speaker attached to the Jetson, save the configuration, and evaluate the same conversation pipeline with server-local USB audio. Headless operation is experimental; this is a conversation workbench, not a robot motion-control framework.
+
+Formerly **Multi-modal AI Studio (MMAS)**. Starting with 0.2.0, the distribution and recommended CLI are `voice-ai-workbench`. The GitHub repository, Python import `multi_modal_ai_studio`, existing settings/session formats, and legacy `multi-modal-ai-studio` CLI remain compatible. See the [upgrade instructions](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/development/releasing.md#upgrading-an-earlier-mmas-installation) before replacing an earlier MMAS wheel.
+
+## What you can do
+
+- **Compare speech services:** Speaches Faster-Whisper/Kokoro, NVIDIA Nemotron ASR/Magpie reference services, and NVIDIA Riva gRPC. REST and Realtime API support and streaming behavior depend on the provider.
+- **Choose an independent LLM:** Connect an OpenAI-compatible Chat Completions endpoint from llama.cpp, vLLM, Ollama, or a hosted service. Change the system prompt to tune your assistant's character and responses.
+- **Tune the conversation:** Discover models and voices, preview a voice, adjust speech speed where supported, and save reusable presets.
+- **Inspect each turn:** View audio waveforms, VAD/ASR, LLM, TTS, barge-in, and CPU/GPU utilization alongside latency measurements. Save sessions with configuration/device snapshots and review or download the session JSON.
+- **Choose your devices:** Use browser microphone/speaker over WebRTC, or Linux server-local USB audio. Camera/VLM support is optional and installed separately.
+
+## Architecture
+
+![Voice AI Workbench architecture (legacy MMAS diagram)](https://raw.githubusercontent.com/NVIDIA-AI-IOT/multi_modal_ai_studio/main/docs/images/MMAS-arch.png)
+
+The workbench connects to model services; it does not bundle their weights or inference runtimes in the Python package.
+
+## Quick Start
+
+### 1. Install the application
+
+Use **Python 3.11+** and a virtual environment. The commands below target 0.2.0; check [GitHub Releases](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/releases) for publication status. Before PyPI publication, install the verified candidate wheel instead of using the package-name command.
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate
+python -m pip install voice-ai-workbench==0.2.0
+voice-ai-workbench --version
+```
+
+On Windows, activate with `.venv\Scripts\activate` instead. On Ubuntu, if virtual-environment creation fails because `venv` or `ensurepip` is missing, install `python3-venv` or the matching version package (for example, `python3.12-venv`).
+
+The wheel includes the Web UI and built-in presets. To open the UI and inspect saved sessions without starting model services:
+
+```bash
+voice-ai-workbench --port 8092
+```
+
+Open `https://localhost:8092`, or `https://JETSON_IP:8092` from another computer. The default certificate is self-signed; verify that you are connecting to your device before accepting it. Microphone access requires HTTPS and browser permission. Stop the foreground application with **Ctrl+C** before restarting it with a preset.
+
+### Recommended quick start on Jetson
+
+The recommended conversation path is **Faster-Whisper tiny.en → Gemma 4 E2B → Kokoro**, using the **Speaches on Jetson — REST** preset. It has been exercised on Jetson Orin Nano and AGX Thor. No separate Ollama installation, Riva server setup, NGC entitlement, or hosted inference key is required.
+
+The backend launch scripts are **repository tools, not part of the wheel**. After the release is published, obtain a matching checkout for those scripts; keep using the pip-installed application:
+
+```bash
+git clone --branch v0.2.0 --depth 1 \
+  https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio.git
+cd multi_modal_ai_studio
+
+# Check the Jetson host, then start ASR/TTS and download the starter models.
+./scripts/speaches_speech.sh doctor
+./scripts/speaches_speech.sh start
+./scripts/speaches_speech.sh verify
+
+# Start text-only Gemma 4 E2B with the appropriate Orin or Thor container.
+./scripts/gemma4_llm.sh start
+./scripts/gemma4_llm.sh verify
+
+# Start the installed application from outside the source checkout.
+cd ..
+voice-ai-workbench --preset speaches-jetson --port 8092
+```
+
+Docker, the NVIDIA container runtime, and a host GPU driver compatible with **CUDA 13 containers** are required for these launchers. Do not assume every JetPack release supports this image. See [Speaches on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_speaches_jetson.md) for prerequisites, model-cache storage, ports, memory considerations, and overrides.
+
+The launcher pins **Speaches 0.9.0-rc.3**, a backend release candidate independent of the workbench's version. **Use REST for the public quick start:** the pinned image's Realtime ASR path has known concurrent-load failures on Nano. A corrected local backend passed follow-up tests but is not yet in that published image; installing this Python package alone does not apply that backend fix. See the [qualification record](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/releases/v0.2.0-qualification.md#realtime-concurrent-load-follow-up--november-5-2026).
+
+### NVIDIA open speech models
+
+To compare **Nemotron 3.5 ASR and Magpie TTS** without setting up a Riva server, run these commands from the matching repository checkout:
+
+```bash
+./scripts/nvidia_open_models_speech.sh build
+./scripts/nvidia_open_models_speech.sh start
+./scripts/nvidia_open_models_speech.sh verify
+```
+
+This launcher manages **only ASR and TTS**, not the LLM. Keep a separately managed OpenAI-compatible LLM running, then select **NVIDIA Speech — REST** or **NVIDIA Speech — Realtime** in the UI. See [NVIDIA open speech models on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_open_models_jetson.md) for exact models, requirements, and streaming limits. Existing Riva users can instead connect their gRPC server using the [Riva setup guide](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_riva.md).
+
+### Optional devices and vision
+
+Install only the features you use:
+
+```bash
+python -m pip install 'voice-ai-workbench[audio]==0.2.0'         # PyAudio devices
+python -m pip install 'voice-ai-workbench[camera]==0.2.0'        # server USB camera
+python -m pip install 'voice-ai-workbench[video-input]==0.2.0'   # video files for VLMs
+python -m pip install 'voice-ai-workbench[webrtc-camera]==0.2.0'
+```
+
+Linux ALSA USB audio uses `arecord`/`aplay` from `alsa-utils`; it does not require the PyAudio extra. PyAudio needs PortAudio system dependencies. See [installation and device setup](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/INSTALL.md).
+
+### Development installation
+
+For source changes, use a separate environment and an editable install instead of the release wheel:
+
+```bash
+git clone https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio.git
+cd multi_modal_ai_studio
+python3 -m venv .venv
+source .venv/bin/activate
+python -m pip install -e .
+```
+
+Alternatively, `./scripts/setup_dev.sh` prepares a development environment. Sample sessions in `mock_sessions/` are available in the repository checkout; run with `--session-dir mock_sessions` to review them. New sessions are stored in `sessions/` relative to the working directory unless `--session-dir` is specified.
+
+## Scope and limitations
+
+- **Streaming depends on the backend.** A Realtime WebSocket API does not guarantee model-native incremental recognition or synthesis. Speaches Faster-Whisper uses VAD-delimited utterances and may not emit partial transcripts.
+- **Early TTS uses text chunks.** “Start speaking before LLM finishes” can reduce waiting, but short chunks affect prosody and shared GPU load can cause playback gaps. Disable it for a baseline quality comparison.
+- **Small devices need memory headroom.** Stop unrelated GPU services and start with the small preset. Cold model loading can delay the first response; swap does not increase GPU memory.
+- **USB/headless validation is narrower than browser validation.** Speaker placement and hardware echo cancellation affect self-transcription and barge-in. Validate your own microphone/speaker setup before robot integration.
+- **Use a trusted development network.** Do not expose the unauthenticated UI or model endpoints directly to the public internet. Functional tests do not substitute for dependency/security review.
+
+## Documentation
+
+| Guide | Description |
+| --- | --- |
+| [Installation](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/INSTALL.md) | Package installation, devices, and troubleshooting |
+| [Speaches on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_speaches_jetson.md) | Recommended Faster-Whisper + Gemma 4 E2B + Kokoro setup |
+| [NVIDIA open speech models](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_open_models_jetson.md) | Nemotron ASR + Magpie TTS reference services |
+| [Speech API backends](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/api_backends.md) | REST, Realtime, and Riva contracts |
+| [Realtime speech](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/realtime_speech.md) | Transcription and response-audio integration |
+| [VLM guide](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/vlm_guide.md) | Vision-language models and frame capture |
+| [Architecture](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/development/architecture.md) | System design and components |
+| [0.2.0 release notes](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/releases/v0.2.0.md) | Changes, compatibility, and validation scope |
+
+## Contributing
+
+Issues, pull requests, and feedback are welcome. Start with the [repository](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio) and describe your device, model services, preset, and reproduction steps.
+
+## License
+
+Apache License 2.0. See [LICENSE](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/LICENSE). Model weights and backend containers have their own licenses and requirements.

voice_ai_workbench-0.2.0/README.md

@@ -0,0 +1,147 @@
+# Voice AI Workbench
+
+**Give your robot a voice. Compare models, shape its character, and measure the conversation.**
+
+![Voice AI Workbench showing a recorded NVIDIA Speech conversation and its latency timeline](https://raw.githubusercontent.com/NVIDIA-AI-IOT/multi_modal_ai_studio/main/docs/images/voice-ai-workbench-nvidia-speech.png)
+
+Recorded on Jetson AGX Thor: Nemotron 3.5 ASR → Gemma 4 E2B → Magpie TTS, using the NVIDIA Speech — Realtime preset. The waveforms, partial transcripts, latency bands, and CPU/GPU traces come from a real session; the internal hostname and IP addresses are omitted from this public screenshot. These timings illustrate one conversation, not a controlled benchmark.
+
+Voice AI Workbench is a developer tool for building and tuning conversational AI. Connect independent speech recognition (ASR), language model (LLM), and text-to-speech (TTS) services; adjust the LLM system prompt; and compare recognition, response latency, playback, and interruptions on a shared timeline. Optional camera input lets you evaluate vision-language models too.
+
+Start with your browser's microphone and speaker for a quick test. Then select a microphone and speaker attached to the Jetson, save the configuration, and evaluate the same conversation pipeline with server-local USB audio. Headless operation is experimental; this is a conversation workbench, not a robot motion-control framework.
+
+Formerly **Multi-modal AI Studio (MMAS)**. Starting with 0.2.0, the distribution and recommended CLI are `voice-ai-workbench`. The GitHub repository, Python import `multi_modal_ai_studio`, existing settings/session formats, and legacy `multi-modal-ai-studio` CLI remain compatible. See the [upgrade instructions](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/development/releasing.md#upgrading-an-earlier-mmas-installation) before replacing an earlier MMAS wheel.
+
+## What you can do
+
+- **Compare speech services:** Speaches Faster-Whisper/Kokoro, NVIDIA Nemotron ASR/Magpie reference services, and NVIDIA Riva gRPC. REST and Realtime API support and streaming behavior depend on the provider.
+- **Choose an independent LLM:** Connect an OpenAI-compatible Chat Completions endpoint from llama.cpp, vLLM, Ollama, or a hosted service. Change the system prompt to tune your assistant's character and responses.
+- **Tune the conversation:** Discover models and voices, preview a voice, adjust speech speed where supported, and save reusable presets.
+- **Inspect each turn:** View audio waveforms, VAD/ASR, LLM, TTS, barge-in, and CPU/GPU utilization alongside latency measurements. Save sessions with configuration/device snapshots and review or download the session JSON.
+- **Choose your devices:** Use browser microphone/speaker over WebRTC, or Linux server-local USB audio. Camera/VLM support is optional and installed separately.
+
+## Architecture
+
+![Voice AI Workbench architecture (legacy MMAS diagram)](https://raw.githubusercontent.com/NVIDIA-AI-IOT/multi_modal_ai_studio/main/docs/images/MMAS-arch.png)
+
+The workbench connects to model services; it does not bundle their weights or inference runtimes in the Python package.
+
+## Quick Start
+
+### 1. Install the application
+
+Use **Python 3.11+** and a virtual environment. The commands below target 0.2.0; check [GitHub Releases](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/releases) for publication status. Before PyPI publication, install the verified candidate wheel instead of using the package-name command.
+
+```bash
+python3 -m venv .venv
+source .venv/bin/activate
+python -m pip install voice-ai-workbench==0.2.0
+voice-ai-workbench --version
+```
+
+On Windows, activate with `.venv\Scripts\activate` instead. On Ubuntu, if virtual-environment creation fails because `venv` or `ensurepip` is missing, install `python3-venv` or the matching version package (for example, `python3.12-venv`).
+
+The wheel includes the Web UI and built-in presets. To open the UI and inspect saved sessions without starting model services:
+
+```bash
+voice-ai-workbench --port 8092
+```
+
+Open `https://localhost:8092`, or `https://JETSON_IP:8092` from another computer. The default certificate is self-signed; verify that you are connecting to your device before accepting it. Microphone access requires HTTPS and browser permission. Stop the foreground application with **Ctrl+C** before restarting it with a preset.
+
+### Recommended quick start on Jetson
+
+The recommended conversation path is **Faster-Whisper tiny.en → Gemma 4 E2B → Kokoro**, using the **Speaches on Jetson — REST** preset. It has been exercised on Jetson Orin Nano and AGX Thor. No separate Ollama installation, Riva server setup, NGC entitlement, or hosted inference key is required.
+
+The backend launch scripts are **repository tools, not part of the wheel**. After the release is published, obtain a matching checkout for those scripts; keep using the pip-installed application:
+
+```bash
+git clone --branch v0.2.0 --depth 1 \
+  https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio.git
+cd multi_modal_ai_studio
+
+# Check the Jetson host, then start ASR/TTS and download the starter models.
+./scripts/speaches_speech.sh doctor
+./scripts/speaches_speech.sh start
+./scripts/speaches_speech.sh verify
+
+# Start text-only Gemma 4 E2B with the appropriate Orin or Thor container.
+./scripts/gemma4_llm.sh start
+./scripts/gemma4_llm.sh verify
+
+# Start the installed application from outside the source checkout.
+cd ..
+voice-ai-workbench --preset speaches-jetson --port 8092
+```
+
+Docker, the NVIDIA container runtime, and a host GPU driver compatible with **CUDA 13 containers** are required for these launchers. Do not assume every JetPack release supports this image. See [Speaches on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_speaches_jetson.md) for prerequisites, model-cache storage, ports, memory considerations, and overrides.
+
+The launcher pins **Speaches 0.9.0-rc.3**, a backend release candidate independent of the workbench's version. **Use REST for the public quick start:** the pinned image's Realtime ASR path has known concurrent-load failures on Nano. A corrected local backend passed follow-up tests but is not yet in that published image; installing this Python package alone does not apply that backend fix. See the [qualification record](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/releases/v0.2.0-qualification.md#realtime-concurrent-load-follow-up--november-5-2026).
+
+### NVIDIA open speech models
+
+To compare **Nemotron 3.5 ASR and Magpie TTS** without setting up a Riva server, run these commands from the matching repository checkout:
+
+```bash
+./scripts/nvidia_open_models_speech.sh build
+./scripts/nvidia_open_models_speech.sh start
+./scripts/nvidia_open_models_speech.sh verify
+```
+
+This launcher manages **only ASR and TTS**, not the LLM. Keep a separately managed OpenAI-compatible LLM running, then select **NVIDIA Speech — REST** or **NVIDIA Speech — Realtime** in the UI. See [NVIDIA open speech models on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_open_models_jetson.md) for exact models, requirements, and streaming limits. Existing Riva users can instead connect their gRPC server using the [Riva setup guide](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_riva.md).
+
+### Optional devices and vision
+
+Install only the features you use:
+
+```bash
+python -m pip install 'voice-ai-workbench[audio]==0.2.0'         # PyAudio devices
+python -m pip install 'voice-ai-workbench[camera]==0.2.0'        # server USB camera
+python -m pip install 'voice-ai-workbench[video-input]==0.2.0'   # video files for VLMs
+python -m pip install 'voice-ai-workbench[webrtc-camera]==0.2.0'
+```
+
+Linux ALSA USB audio uses `arecord`/`aplay` from `alsa-utils`; it does not require the PyAudio extra. PyAudio needs PortAudio system dependencies. See [installation and device setup](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/INSTALL.md).
+
+### Development installation
+
+For source changes, use a separate environment and an editable install instead of the release wheel:
+
+```bash
+git clone https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio.git
+cd multi_modal_ai_studio
+python3 -m venv .venv
+source .venv/bin/activate
+python -m pip install -e .
+```
+
+Alternatively, `./scripts/setup_dev.sh` prepares a development environment. Sample sessions in `mock_sessions/` are available in the repository checkout; run with `--session-dir mock_sessions` to review them. New sessions are stored in `sessions/` relative to the working directory unless `--session-dir` is specified.
+
+## Scope and limitations
+
+- **Streaming depends on the backend.** A Realtime WebSocket API does not guarantee model-native incremental recognition or synthesis. Speaches Faster-Whisper uses VAD-delimited utterances and may not emit partial transcripts.
+- **Early TTS uses text chunks.** “Start speaking before LLM finishes” can reduce waiting, but short chunks affect prosody and shared GPU load can cause playback gaps. Disable it for a baseline quality comparison.
+- **Small devices need memory headroom.** Stop unrelated GPU services and start with the small preset. Cold model loading can delay the first response; swap does not increase GPU memory.
+- **USB/headless validation is narrower than browser validation.** Speaker placement and hardware echo cancellation affect self-transcription and barge-in. Validate your own microphone/speaker setup before robot integration.
+- **Use a trusted development network.** Do not expose the unauthenticated UI or model endpoints directly to the public internet. Functional tests do not substitute for dependency/security review.
+
+## Documentation
+
+| Guide | Description |
+| --- | --- |
+| [Installation](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/INSTALL.md) | Package installation, devices, and troubleshooting |
+| [Speaches on Jetson](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_speaches_jetson.md) | Recommended Faster-Whisper + Gemma 4 E2B + Kokoro setup |
+| [NVIDIA open speech models](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/setup_open_models_jetson.md) | Nemotron ASR + Magpie TTS reference services |
+| [Speech API backends](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/api_backends.md) | REST, Realtime, and Riva contracts |
+| [Realtime speech](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/realtime_speech.md) | Transcription and response-audio integration |
+| [VLM guide](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/vlm_guide.md) | Vision-language models and frame capture |
+| [Architecture](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/development/architecture.md) | System design and components |
+| [0.2.0 release notes](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/docs/releases/v0.2.0.md) | Changes, compatibility, and validation scope |
+
+## Contributing
+
+Issues, pull requests, and feedback are welcome. Start with the [repository](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio) and describe your device, model services, preset, and reproduction steps.
+
+## License
+
+Apache License 2.0. See [LICENSE](https://github.com/NVIDIA-AI-IOT/multi_modal_ai_studio/blob/main/LICENSE). Model weights and backend containers have their own licenses and requirements.

voice_ai_workbench-0.2.0/presets/cosmos-reason.yaml

@@ -0,0 +1,67 @@
+# SPDX-FileCopyrightText: Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+# SPDX-License-Identifier: Apache-2.0
+#
+# Cosmos-Reason2 VLM served by vLLM — vision + chain-of-thought reasoning.
+# Flagship Jetson configuration: Riva ASR → vLLM (Cosmos-Reason2) → Riva TTS.
+#
+# Prerequisites:
+#   - Riva ASR/TTS server on localhost:50051
+#   - vLLM serving Cosmos-Reason2 on localhost:8010
+
+name: "Cosmos-Reason2 VLM (vLLM)"
+description: "Riva ASR/TTS + Cosmos-Reason2 via vLLM with vision and reasoning"
+app_version: "0.1.0"
+
+asr:
+  scheme: riva
+  server: localhost:50051
+  model: parakeet-1.1b-en-US-asr-streaming-silero-vad-sortformer
+  language: en-US
+  vad_start_threshold: 0.5
+  vad_stop_threshold: 0.3
+  speech_pad_ms: 600
+  speech_timeout_ms: 1200
+
+llm:
+  scheme: openai
+  api_base: http://localhost:8010/v1
+  model: nvidia/cosmos-reason2-8b-fp8
+  temperature: 0.3
+  max_tokens: 512
+  history_turns: 0
+  include_conversation_history: false
+  system_prompt: >-
+    You are a helpful voice AI assistant.
+    Plain text only, no markdown, no bullet points, no emojis.
+  enable_vision: true
+  vision_system_prompt: >-
+    You are a helpful voice and vision assistant.
+    Give ONE short sentence answers only. Be direct.
+    Plain text only, no markdown, no bullet points, no emojis.
+  vision_frames: 30
+  vision_detail: auto
+  vision_quality: 0.8
+  vision_max_width: 768
+  vision_buffer_fps: 5.0
+  vision_video_encode: true
+  vision_api_format: openai
+  vision_omit_last_assistant: true
+  enable_reasoning_through_user_prompt: true
+
+tts:
+  scheme: riva
+  server: localhost:50051
+  voice: ""
+  sample_rate: 22050
+  stream_tts: true
+
+devices:
+  video_source: browser
+  audio_input_source: browser
+  audio_output_source: browser
+
+app:
+  barge_in_enabled: true
+  timeline_position: right
+  session_auto_save: true
+  session_output_dir: ./sessions