feat(voice-clone-tune): integrate Inno tuner with web ui tab by @remsky in #534 refactor/normalization-classes by @remsky in #528 feat(normalizer): dashes, emoji, phone numbers, all-caps, plural acronyms; web normalization menu and graded voice list (#249, #259, #353) by @remsky in #529 fix(streaming): close audio writer and generator on client disconnect (#288, #337) by @remsky in #533 fix: prese…
Curated summary: CHANGELOG.md What's Changed fix: honour default_voice_code in _split_multi_voice by @Christian-Sidak in #515 Fix/phonemes default voice code by @remsky in #516 feat(api, web ui): include voice grades on the voice listing endpoint by @remsky in #517 Auto unload on timeout by @brycehenson in #506 fix(chunking): replace regex sentence splitter with UAX #29 segmentation (#308, adapts…
What's Changed feat(time-to-first-audio): faster streaming responses via lazy phonemization in get_sentence_info by @remsky in #513 fix(schema, pause overflow): caps total pause duration, input length, volume_multiplier, validates lang_code by @remsky in #511 guards malicious OOM risk on public endpoints: GHSA-f64g-9jmv-pm22 submitted via @sshpie fix(voices): combine endpoint accepts weighted synt…
Curated summary: CHANGELOG.md What's Changed feat(multi-speaker): inline multi-speaker support, voice alias mappings, read-along web ui by @remsky with @radzrader in #502 feat(ssml, webui): add experimental SSML support, web UI improvements, documentation by @remsky in #503 build(docker): bytecode compile for ~40% faster startup, entrypoint fix for unraid/optimizations, lockfile alignment by @rems…
master state before ssml branch merge
Curated summary: CHANGELOG.md What's Changed fix: flac loses its tail at finalize (#497) by @remsky in #499 build: bump fastapi + pin starlette>=1.3.1 to cover CVE's (DoS via FileResponse) by @bunlongheng in #500 ci(docker): switch build cache from GHA to GHCR registry to allow public... by @remsky in #501 New Contributors @bunlongheng made their first contribution in #500 Full Changelog: v0.7.1..…
Curated summary: CHANGELOG.md What's Changed fix(audio): improve error handling for aborted audio streams by @yooudeek in #490 feat(web ui): refactor; downloaded file naming, better use of space, improved responsiveness and styling by @remsky in #495 New Contributors @yooudeek made their first contribution in #490 Full Changelog: v0.7.0...v0.7.1
Curated summary: CHANGELOG.md What's Changed build(docker): optimizations and python-3.12 migration complete on pre-built images by @remsky in #491 and #482 Image sizes (compressed, amd64): CPU: 1.66 -> 1.56 GB GPU: 6.81 -> 4.68 GB (-31%) GPU cu128: 8.11 -> 5.23 GB (-36%) ROCm: 13.08 -> 13.36 GB (model now bakes at build, unchanged base) fix(download_model): release artifact checksums, temp-dir av…
Curated summary: CHANGELOG.md What's Changed fix(mappings): update voice mappings for consistency, add missing aud… by @remsky in #480 Prevent POST /dev/unload behind an opt-in setting by @sanjaygulati13 in #483 Fix missing word timestamps for non-English (espeak) voices by @mlevinmail in #484 fix(debug-endpoints): gate /debug/* introspection routes behind ENABL… by @remsky in #486 New Contributor…
Curated summary: CHANGELOG.md What's Changed fix(webui): bounded buffer to resolve 10 minute crashout, player stat… by @remsky in #471 feat(api): add POST /dev/unload to release model from GPU VRAM by @sageframe-no-kaji in #474 fix(audio): implement audio source swapping for stream playback/seeking by @remsky in #475 New Contributors @sageframe-no-kaji made their first contribution in #474 Full Ch…
Curated summary: CHANGELOG.md What's Changed feat(gpu): adds clear cu128 opt-in via compose, setup for Blackwell /… by @remsky in #466 fix(audio): drop WAV close()-time junk trailer that caused chunk-end click #463 by @remsky in #467 chore(docker): add oci metadata, copy VERSION for web ui, prebuilt image for tts-client tests by @remsky in #468 fix(rocm): persist MIOpen cache and add FIND_MODE=2 d…
Changes to simplify streaming/async inference pathways still somewhat in progress. WebUI added as a lighter-weight alternative to the Gradio UI More of the configuration variables are exposed, temporary file management settings Added new debug endpoints for system and storage information (threads, sessions, etc) Significant restructuring towards concurrency, decoupling inference workflows, more fl…
What's Changed docs: added a note for Apple Silicon users regarding GPU build by @FotieMConstant in #232 converted CRLF ending lines to LF ones in api/src/structures/custom_responses.py by @blakkd in #199 Fixes and Features by @fireblade2534 in #235 Add Direct Windows support scripts by @kimnzl in #241 Added support for MPS on Apple silicon by @rampadc in #233 use local js file instead of the unpk…
What's Changed Fix custom phenomes and make them much more robust by @fireblade2534 in #350 Update PyTorch to 2.7.1+cu128 for RTX 50 Series GPU Support by @MiggiV2 in #354 Smaller docker image by @faltiska in #361 Use torch 2.8 for compatibility with 50 series by @garciadias in #390 feat: add configurable logging level with environment variable support by @ryan-steed-usa in #404 Fix(cpu/docker): i…
What's Changed docs: added a note for Apple Silicon users regarding GPU build by @FotieMConstant in #232 converted CRLF ending lines to LF ones in api/src/structures/custom_responses.py by @blakkd in #199 Fixes and Features by @fireblade2534 in #235 Add Direct Windows support scripts by @kimnzl in #241 Added support for MPS on Apple silicon by @rampadc in #233 use local js file instead of the unpk…
What's Changed Disable --reload on unicorn/fastapi to avoid pegging a CPU core by @randombk in #171 Add a .gitattributes by @fireblade2534 in #186 Normalization changes by @fireblade2534 in #179 Streaming word timestamps by @fireblade2534 in #173 Fix low quality because audio was being encoded at a lower bitrate by @fireblade2534 in #207 New Contributors @randombk made their first contribution in…
Fixes speak not engaging reliably on the CPU image as a fallback audio quality bumped up by adjusting compression settings, bug with webui format selection advanced normalization settings added @fireblade2534 What's Changed Add Helm chart by @zucher in #157 #162 fixed a bunch of stuff by @fireblade2534 in #152 added settings based override of default lang_code by @Krurst in #155 docs update @eltoc…
What's Changed adjustment to improve compatibility with espeak-loader dependency on misaki #127 added v1/models dummy endpoint for compatibility #144 fixed issue with duplicates captions, swapping to a stream on audio + tempfile download at completion for caption files #139 fixed some problems in the build system and model download system by @fireblade2534 in #131 Full Changelog: v0.2.0...v0.2.1
Complete Model Overhaul: Upgraded to Kokoro v1.0 model architecture, deprecated V0.19 support Integration with hexgrad/kokoro and hexgrad/misaki packages Pre-installed all multi-language support from Misaki: English (en), Japanese (ja), Korean (ko), Chinese (zh), Vietnamese (vi) Note: This will likely controlled via env variable in upcoming versions All voice packs included for supported languages…