Skip to content

audio.cpp Windows portable v0.5.0

Choose a tag to compare

View all tags
@kigner kigner released this 08 Aug 12:55
· 1 commit to v0.5.0 since this release
v0.5.0-windows-prebuilt
18eaad2

audio.cpp Windows portable v0.5.0

This stable release combines audio.cpp 0.5 and the upstream changes through
3ba02c9 with the local Windows WebUI and updater distribution.

Highlights

  • Expanded the shared runtime to 44 model families, including DramaBox,
    Confucius4-TTS, RVC, BS-RoFormer, GLM-TTS, Kroko ASR, Parakeet-TDT,
    Inflect v2, and Fun-ASR-Nano.
  • Added resilient model downloads with bounded retries, byte-range resume,
    safe restart when a CDN ignores ranges, and completed-file handling.
  • Simplified the WebUI download flow to one click while retaining disk-space
    blockers and visible progress/error reporting.
  • Added WebUI GGUF conversion, status, loading, and deletion for supported
    safetensors model packages.
  • Hardened Qwen3-TTS long-text generation with Unicode-safe chunking,
    max-token-aware defaults, and explicit truncation detection.
  • Changed Parakeet-TDT buffered streaming output to append-only transcript
    deltas and prepared its streaming input as mono 16 kHz PCM.
  • Made the forced-aligner transcript input available on initial render and
    added the active app version to the WebUI footer.
  • Aligned the portable app package with model manager v2 and the current
    model-spec catalog.

Windows builds

  • CPU core: AVX2
  • CUDA core architectures:
    61-real;75-real;86-real;89-real;120a-real;90-virtual
  • Updated executables: audiocpp_cli, audiocpp_server, and audiocpp_gguf
  • Python environment changes: none
  • CUDA runtime DLL changes: none
  • Updater version: 1.1.1

Updating

Run update.bat from an existing updater-enabled Windows portable
installation. The stable channel updates directly to v0.5.0; no intermediate
release is required.

  • Supported source versions: 0.2.0 and later
  • Verified direct-upgrade matrix: 0.2.0, 0.2.1, 0.2.2, 0.3.0,
    0.4.0, 0.4.1, and 0.4.2 to 0.5.0
  • Updated components: app, core-cpu, and core-cuda

Portable releases older than 0.2.0 did not include the current updater
contract. Those installations require the one-time updater bootstrap before
they can use the stable channel.

The updater preserves downloaded models, custom voices, generated output, API
keys, logs, UI language preferences, and SpeakType configuration. If
validation fails, managed files are restored from _update\backup\.

Validation

  • CPU and CUDA release binaries are rebuilt from the v0.5.0 source commit.
  • WebUI, model-manager, catalog, syntax, and updater acceptance tests pass.
  • Every updater-supported historical version is validated as a direct source
    for the signed v0.5.0 manifest.
  • Release manifests, checksums, component indexes, and published assets are
    verified with the bundled minisign public key.
Assets 11
Loading