Documentation¶
The core documentation set is:
README.mddocs/getting-started.mddocs/api.mddocs/architecture.mddocs/faq.md
Use those five files for the package overview, first run, supported API, implementation map, cache/history reset, and common troubleshooting.
If you’re using AbstractVoice as part of the AbstractFramework ecosystem
(AbstractCore / AbstractRuntime), see the integration notes in README.md and
the “Integrations” section in docs/api.md. In that mode, AbstractCore owns
the server/API surface; AbstractVoice is the capability plugin that backs
voice/audio operations such as /v1/audio/speech and
/v1/audio/transcriptions.
For a browser smoke test, install abstractvoice[web] and run
OPENAI_API_KEY=... abstractvoice web. That web UI is a small VoiceManager example with
assistant/user voice selection, browser voice cloning from uploaded or recorded
reference audio, and an example-only local LLM dialogue bridge, not the
production API server. Compose extras such as abstractvoice[web,local] or
abstractvoice[web,omnivoice] when you want local engines in the same install.
User-facing docs¶
- FAQ:
docs/faq.md - Known issues:
docs/known-issues.md - Install & troubleshooting:
docs/installation.md - REPL guide (voice assistant):
docs/repl_guide.md - Multilingual:
docs/multilingual.md - Voice/model management:
docs/model-management.md - Dependencies & model sources:
docs/dependencies.md - Licensing notes for models/voices:
docs/voices-and-licenses.md - Benchmarks (TTS-only):
docs/benchmark.md
For integrators¶
- Stable API surface:
docs/api.md - Architecture (implementation map):
docs/architecture.md - ADRs (design decisions):
docs/adr/
Project¶
- Changelog:
CHANGELOG.md - Contributing:
CONTRIBUTING.md - Bug reports:
.github/ISSUE_TEMPLATE/bug_report.yml - Security:
SECURITY.md - Acknowledgments:
ACKNOWLEDGMENTS.md - License:
LICENSE
Internal / historical (not an API contract)¶
These documents are useful for maintainers, but may describe past work or future plans:
docs/backlog/(planning + task notes)docs/reports/(status and implementation reports)docs/voice_cloning_2026.md(research notes; not a promise of product behavior)