Voice text services

English reading guide. The complete Chinese specification remains the authoritative source for exact fields, examples, implementation status, and historical details. This page is a concise guide, not a full translation.

Use the supported text-service boundary

• Declare and invoke the actual versioned service capability. A text request and voice recognition have different inputs and authorization conditions.

• Preserve the user's actual request, current input scope, and selected service context. Interface language does not automatically change generated output language.

• Handle unavailable service, refusal, cancellation, and provider errors using the real contract.

• The source defines precise request and response fields and the current capability boundary; use it when implementing a consumer.

Continue reading

• Complete Chinese specification
• Documentation overview
• Plugin development
• API reference
• Packaging and submission

Source: docs/voice-request-text.md.