v0.1.0

Osmantic/ODSv0.1.0Aug 31, 2026by bilawalsidhu

AI Summary

This is the initial versioned release of NeMo-Speech.cpp, introducing a comprehensive suite of speech processing capabilities including recognition, diarization, synthesis, and translation across multiple backends and platforms.

Key Highlights

  • Support for diverse speech recognition models (Nemotron 3.5, Parakeet) and synthesis (MagpieTTS, NanoCodec)
  • Native C API and OpenAI-compatible HTTP API for developer integration
  • Cross-platform backend support including CPU, CUDA, Vulkan, and Metal
  • Unified CLI for managed model discovery, downloads, and caching

New Features

  • Speech recognition with Nemotron 3.5, Nemotron Speech Streaming, Parakeet TDT, and Parakeet CTC
  • Speaker diarization with Streaming Sortformer v2
  • Speech synthesis with MagpieTTS and NanoCodec
  • Text and speech translation with Riva Translate 4B Instruct v2
  • Subtitles, VAD, endpointing, punctuation, and capitalization
  • Unified nemo-speech CLI with managed model discovery, downloads, and caching
  • Local HTTP API with OpenAI-compatible transcription and realtime WebSocket transcription
  • Native C API, shared libraries, and exported CMake package
  • CPU, CUDA, Vulkan, and Metal backend support

Full Release Notes

The first formal release: one-click install, a keyless boot that lands on a real satellite globe, and in-app key management.

**Get started in one click → [Quick Start](https://github.com/bilawalsidhu/gods-eye-view#-quick-start)** — no keys, no signup, no config file.

## Highlights

- **One-click install** via Pinokio. Keyless boot lands on live Esri World Imagery with keyless terrain; OSM takes over automatically if Esri is unreachable, and the globe continues without terrain if its source is unavailable.
- **Provider Settings** (the POWER UP panel): add, replace, or remove API keys inside the app. Credential files are made owner-only before any secret is written — verified on macOS and Windows — and keys configured outside the panel are shown read-only. The panel answers only local, unproxied requests and disables itself whenever the server is shared.
- **Keyless capability responses**: the optional HUD summary and place-search endpoints return a deliberate "not configured" success instead of errors, without consuming rate-limit quota.
- **Aircraft-identity voice answers** explicitly cover operator, type, and route, and say so plainly when enrichment is unavailable.
- **README rewritten keyless-first** around the provider ladder (zero-key Esri → free Cesium ion for eligible personal/non-commercial use → direct, billing-enabled Google), with the walkthrough video and corrected media paths. The changelog gains a concise 0.1.0 section with pre-release history retained below it.
- **CI arrives with this PR** (unit suite + build on Node 24/26 + Windows onboarding job).

## Validation (local, on this exact tree)

| Gate | Result |
|---|---|
| Unit suite | 2,672 pass / 0 fail |
| Build | clean |
| Tracking regression (live dev server) | 108/108 |
| Map-source tray QA (keyless fallback) | pass |
| Setup doctor | ready |

## Community

- **@ethanstoner** — thank you for the repo's first merged contributions: #81 (line-ending normalization via `.gitattributes`, which un-breaks the test suite on Windows clones) and #83 (keeping `node:fs` out of the browser-built data modules, plus the boundary scan test that guards it). Both authored commits remain intact in v0.1.0.
- **@tdkxygonjhif-blip** — for requesting a keyless provider option (#64), and **@C-EO** for suggesting Esri as the route. That is exactly what shipped as the zero-key default.
- **@xsoft** — special credit for researching and proving the Cesium ion route to Google Photorealistic 3D from Europe/EEA (#71). That work is now a supported, documented path.
- **@asabino2** — for reporting the Google authentication failure (#59) and making the need for a documented alternative route obvious.
- **@Anoymoushackz** and **@eabase** — for surfacing the setup, billing, and privacy friction (#69, #74) addressed by keyless boot, the ion route, and Provider Settings.

Thanks again to everyone in #59, #64, #69, #71, and #74 — I'll share the release in those threads once it's live.