Crate gestura_core_streaming 

Streaming response types, config, and cancellation support for Gestura.

gestura-core-streaming provides the shared building blocks for incremental LLM response delivery across the workspace. It defines the streaming event model surfaced to UIs, the decoupled streaming configuration types used at runtime, and the cancellation registry that coordinates interruption of in-flight streams.

Responsibilities

• streaming: core streaming events, provider-specific streaming helpers, token-usage updates, tool-call chunks, shell output chunks, and status notifications used by CLI/GUI frontends
• config: a portable StreamingConfig model derived from application configuration without tying the streaming crate directly to AppConfig
• cancellation: cooperative cancellation-token registry for active streams

Architecture boundary

This crate owns the streaming protocol and runtime primitives. Higher-level orchestration—such as deciding when to start a stream, how to bridge it into sessions, and how to mix streaming with tool execution—remains in gestura-core.

Stable import paths

Most consumers should import through the facade re-exports:

• gestura_core::streaming::*
• gestura_core::stream_cancellation::*

Why the config is separate

StreamingConfig mirrors only the subset of provider configuration needed for streaming. This keeps the crate portable, easier to test, and less tightly coupled to the larger application configuration surface.

Re-exports

pub use cancellation::STREAM_CANCELLATIONS;

pub use cancellation::StreamCancellationRegistry;

pub use config::AnthropicProviderConfig;

pub use config::GeminiProviderConfig;

pub use config::GrokProviderConfig;

pub use config::OllamaProviderConfig;

pub use config::OpenAiProviderConfig;

pub use config::StreamingConfig;

pub use streaming::*;

Modules

cancellation

Streaming cancellation registry.

config

Streaming-specific configuration types.

streaming

Streaming LLM provider support for Gestura