Update

[usabarashi]

No description provided.

[gemini-code-assist]

Summary of Changes

Hello @usabarashi, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request focuses on enhancing the robustness, modularity, and user-friendliness of the VOICEVOX CLI, particularly its Model Context Protocol (MCP) server. Key improvements include a complete refactoring of the MCP server's internal architecture to support cancellable tool execution, adopting XDG Base Directory standards for configuration, and streamlining the development and CI environments. These changes aim to provide a more stable and configurable experience for users and developers alike, while also updating underlying dependencies for better performance and security.

Highlights

• MCP Server Refactoring and Cancellation Support: The Model Context Protocol (MCP) server has undergone a significant refactoring, moving core logic from monolithic handlers into dedicated modules for protocol handling (protocol.rs), request management and cancellation (requests.rs), and tool definitions/execution (tools.rs). This includes implementing cancellable tool execution, particularly for audio playback, allowing clients to stop ongoing operations.
• XDG Base Directory Compliance for Instructions: The MCP server's instruction loading mechanism now adheres to the XDG Base Directory Specification. This provides a clear, prioritized order for locating instruction files (VOICEVOX.md) from environment variables, user configuration directories, executable paths, and the current working directory, enhancing configurability and user experience.
• Nix Environment and CI Script Simplification: The Nix flake configuration has been streamlined, removing Serena-specific wrappers and simplifying the development shell setup. The CI script (scripts/ci.sh) has also been improved with a new helper function for Nix environment commands and refined build artifact verification.
• Dependency Updates and Codebase Cleanup: Numerous Rust crate dependencies, including anyhow, clap, serde, thiserror, and tempfile, have been updated to their latest versions. Additionally, several modules related to configuration loading (src/config.rs), setup (src/setup.rs), and verbose logging have been removed or simplified, leading to a cleaner and more maintainable codebase.
• Updated Documentation and Installation: Documentation files (CLAUDE.md, README.md, docs/mcp-usage.md) have been updated to reflect the new instruction loading priority, simplified Nix environment, and revised quick start installation methods for pre-built binaries.

Ignored Files

• Ignored by pattern: .github/workflows/** (4)
  • .github/workflows/branch-restrictions.yml
  • .github/workflows/ci.yml
  • .github/workflows/release.yml
  • .github/workflows/update-flake.yml

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

---

Code refactored, clean and new, Modules shine, a clearer view. No more old paths, now XDG, Instructions flow, for all to see. Dependencies, updated, true.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[Copilot]

Pull Request Overview

Refactors MCP server architecture with protocol separation, introduces cancellation support for tool executions, simplifies configuration and setup logic, and removes numerous logging/output and legacy setup/config code. Key changes:

• MCP subsystem refactored (new protocol, requests, tools) with async execution and cancellation; JSON-RPC / instruction loading logic reworked (XDG-based).
• Removed configuration load/save and setup/model bootstrap code; replaced Config::load usage with Config::default; reduced logging and error reporting across multiple modules.
• Updated build/packaging (flake.nix simplification, CI scripts), added temp file audio handling and cancellation-capable playback.

Reviewed Changes

Copilot reviewed 29 out of 32 changed files in this pull request and generated 7 comments.

Show a summary per file

File	Description
src/voice.rs	Suppressed model load/unload/logging errors during mapping to simplify output.
src/synthesis/streaming.rs	Replaced persisted config loading with default config instantiation.
src/setup.rs	Removed first-run model setup helper.
src/paths.rs	Removed models dir helper and silenced unexpected filename warning.
src/mcp/types.rs	Removed old MCP type definitions (replaced by protocol.rs).
src/mcp/tools.rs	Added full MCP tool framework with cancellation, playback variants, validation, tests.
src/mcp/server.rs	Rewrote server loop with request routing, response channel, cancellation integration.
src/mcp/requests.rs	New active request tracking & async execution management.
src/mcp/protocol.rs	New MCP protocol definitions, instruction loading (XDG), request/notification handling.
src/mcp/mod.rs	Updated module exports to new structure.
src/lib.rs	Removed setup module export.
src/daemon/server.rs	Removed verbose logging; simplified model load/unload and error handling.
src/daemon/process.rs	Suppressed warning logging on process discovery failure.
src/config.rs	Removed file-based config load/save; left only struct defaults.
src/client/audio.rs	Added robust temp WAV file creation; refactored playback logic.
scripts/ci.sh	Added helper function; streamlined artifact and tool invocation.
flake.nix	Simplified flake, removed auxiliary tooling scripts, adjusted inputs and dev shell.
docs/mcp-usage.md	Updated instruction loading documentation to XDG-based VOICEVOX.md.
VOICEVOX.md	New instruction file replacing INSTRUCTIONS.md.
README.md	Updated installation instructions for non-Nix distribution.
INSTRUCTIONS.md	Removed legacy instruction file.
Cargo.toml	Dependency version adjustments; removed unused crates.
CLAUDE.md	Updated instruction loading priority documentation.
.mcp.json	Updated Serena MCP server invocation method/environment.
.github/workflows/*	Bumped action versions; minor permissions addition.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[Copilot]

The input URL uses 'nixpkgs/nixpkgs-unstable' which is not a standard ref; the usual channel is 'nixos-unstable' (i.e. github:NixOS/nixpkgs/nixos-unstable) or a rev/ref pair. This will likely fail to resolve; change to 'github:NixOS/nixpkgs/nixos-unstable' unless you intended a different ref.

Suggested change

	nixpkgs.url = "github:NixOS/nixpkgs/nixpkgs-unstable";
	nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";

[Copilot]

Using spawn_blocking + runtime_handle.block_on to run async code introduces unnecessary indirection and can reduce scalability. Prefer spawning the async task directly with tokio::spawn (move async { ... }) and await the tool execution, unless the code inside is truly CPU-bound/blocking.

Suggested change

	let runtime_handle = tokio::runtime::Handle::current();
	tokio::task::spawn_blocking(move || {
	// Use current runtime handle instead of creating a new one
	runtime_handle.block_on(async move {
	let result =
	tools::execute_tool_request(&tool_name, arguments, Some(abort_rx)).await;
	// Clean up the request from active list
	active_requests.complete(&request_id).await;
	// Send response
	let response = match result {
	Ok(tool_result) => match serde_json::to_value(tool_result) {
	Ok(value) => JsonRpcResponse::success(id, value),
	Err(_) => JsonRpcResponse::error(
	id,
	INTERNAL_ERROR,
	"Failed to serialize response".to_string(),
	),
	},
	Err(e) => {
	let error_result = ToolCallResult {
	content: vec![ToolContent {
	content_type: "text".to_string(),
	text: format!("Tool execution error: {e}"),
	}],
	is_error: Some(true),
	};
	match serde_json::to_value(error_result) {
	Ok(value) => JsonRpcResponse::success(id, value),
	Err(_) => JsonRpcResponse::error(
	id,
	INTERNAL_ERROR,
	"Failed to serialize error response".to_string(),
	),
	}
	}
	};
	// Send response via channel
	let _ = active_requests.response_sender.send(response);
	})
	tokio::spawn(async move {
	let result =
	tools::execute_tool_request(&tool_name, arguments, Some(abort_rx)).await;
	// Clean up the request from active list
	active_requests.complete(&request_id).await;
	// Send response
	let response = match result {
	Ok(tool_result) => match serde_json::to_value(tool_result) {
	Ok(value) => JsonRpcResponse::success(id, value),
	Err(_) => JsonRpcResponse::error(
	id,
	INTERNAL_ERROR,
	"Failed to serialize response".to_string(),
	),
	},
	Err(e) => {
	let error_result = ToolCallResult {
	content: vec![ToolContent {
	content_type: "text".to_string(),
	text: format!("Tool execution error: {e}"),
	}],
	is_error: Some(true),
	};
	match serde_json::to_value(error_result) {
	Ok(value) => JsonRpcResponse::success(id, value),
	Err(_) => JsonRpcResponse::error(
	id,
	INTERNAL_ERROR,
	"Failed to serialize error response".to_string(),
	),
	}
	}
	};
	// Send response via channel
	let _ = active_requests.response_sender.send(response);

[Copilot]

Previously this path logged model load failures; now errors are silently ignored, making it harder to diagnose missing styles/models. Consider at least emitting a trace/debug log or aggregating failures for later reporting.

[Copilot]

The assigned variable '_expected_patterns' is never used; this entire block becomes dead code after removing the warning log. Remove it (or reintroduce a log) to avoid misleading future readers.

Suggested change

	let _expected_patterns = if cfg!(target_os = "macos") {
	"libonnxruntime.dylib or libvoicevox_onnxruntime.*.dylib"
	} else if cfg!(target_os = "linux") {
	"libonnxruntime.so or libvoicevox_onnxruntime.*.so"
	} else {
	"onnxruntime.dll, libonnxruntime.dll, or libvoicevox_onnxruntime.*.dll"
	};
	// Invalid ORT_DYLIB_PATH filename; expected platform-specific patterns.

[Copilot]

Replacing Config::load().unwrap_or_default() with Config::default() removes any persisted/user configuration capability that may have existed; this is a silent behavior change for callers expecting overrides. If intentional, document the deprecation or keep a fall-back attempt to load while tolerating failure.

Suggested change

	let config = Config::default();
	let config = Config::load().unwrap_or_default();

[Copilot]

Deserialization errors now cause a silent break without logging, making it difficult to diagnose protocol incompatibilities or corrupted frames. Consider logging at least a concise warning before breaking.

Suggested change

	Err(_) => {
	Err(err) => {
	eprintln!("Warning: Failed to deserialize DaemonRequest: {}", err);

[Copilot]

Swallowing the underlying error when finding daemon processes removes potentially useful diagnostic information (e.g., pgrep not present). Retain minimal logging or map specific errors if noisy output was the concern.

Suggested change

	Err(_) => 0,
	Err(e) => {
	eprintln!("Failed to find daemon processes: {}", e);
	0
	},