From fedfb51e89c4ee5d3dfabee354b5524ed0ffdeb8 Mon Sep 17 00:00:00 2001
From: patrick <patrick@rtk-ai.app>
Date: Thu, 14 May 2026 22:31:52 +0200
Subject: [PATCH 1/2] perf: compress embedded espeak-ng-data with rust-embed

Swap include_dir for rust-embed with the 'compression' feature. The
embedded files were uncompressed phoneme dictionaries (~15 MB plain
text/binary), DEFLATE typically gives 4-5x on this content. Expected
release binary drop: ~50 MB to ~32 MB, tarball ~22 MB to ~13 MB.

UX unchanged: first piper run still extracts to
~/.config/vox/piper/espeak-ng-data and sets PIPER_ESPEAKNG_DATA_DIRECTORY.
The sentinel file logic short-circuits on subsequent runs.
---
 Cargo.toml           |  2 +-
 src/backend/piper.rs | 27 ++++++++++++++++++---------
 2 files changed, 19 insertions(+), 10 deletions(-)

diff --git a/Cargo.toml b/Cargo.toml
index ad7d4de..2aa0dee 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -29,7 +29,7 @@ rodio = "0.20"
 hound = "3"
 qwen3-tts = { git = "https://github.com/TrevorS/qwen3-tts-rs", features = ["hub"], default-features = false }
 piper-rs = { git = "https://github.com/thewh1teagle/piper-rs" }
-include_dir = "0.7"
+rust-embed = { version = "8", features = ["compression", "interpolate-folder-path"] }
 kokoro-tts = { version = "0.3", optional = true }
 toml = "0.8"
 ratatui = "0.29"
diff --git a/src/backend/piper.rs b/src/backend/piper.rs
index 2a673ac..b031039 100644
--- a/src/backend/piper.rs
+++ b/src/backend/piper.rs
@@ -7,8 +7,8 @@ use std::path::PathBuf;
 use std::sync::{Mutex, OnceLock};
 
 use anyhow::{Context, Result};
-use include_dir::{Dir, include_dir};
 use piper_rs::Piper;
+use rust_embed::RustEmbed;
 
 use super::{SpeakOptions, TtsBackend};
 use crate::config;
@@ -19,11 +19,14 @@ pub struct PiperBackend;
 /// Reloads when language changes (different ONNX model per language).
 static MODEL: Mutex<Option<(String, Piper)>> = Mutex::new(None);
 
-/// espeak-ng-data embedded at build time (staged by build.rs into OUT_DIR).
-/// Needed because the espeak-ng library statically linked into vox has a
-/// hard-coded data path from the CI builder that does not exist on user
-/// machines. We extract this once and point espeak-rs at the result.
-static ESPEAK_DATA: Dir<'_> = include_dir!("$OUT_DIR/espeak-ng-data");
+/// espeak-ng-data embedded (and DEFLATE-compressed) at build time. Staged
+/// into OUT_DIR by build.rs. Needed because the espeak-ng library statically
+/// linked into vox has a hard-coded data path from the CI builder that does
+/// not exist on user machines. We extract this once and point espeak-rs at
+/// the result.
+#[derive(RustEmbed)]
+#[folder = "$OUT_DIR/espeak-ng-data"]
+struct EspeakData;
 
 static ESPEAK_DATA_INIT: OnceLock<Result<PathBuf, String>> = OnceLock::new();
 
@@ -44,9 +47,15 @@ fn ensure_espeak_data() -> Result<()> {
                 std::fs::remove_dir_all(&data_dir).map_err(|e| e.to_string())?;
             }
             std::fs::create_dir_all(&data_dir).map_err(|e| e.to_string())?;
-            ESPEAK_DATA
-                .extract(&data_dir)
-                .map_err(|e| format!("failed to extract espeak-ng-data: {e}"))?;
+            for path in EspeakData::iter() {
+                let file = EspeakData::get(&path)
+                    .ok_or_else(|| format!("embedded espeak-ng-data entry vanished: {path}"))?;
+                let target = data_dir.join(path.as_ref());
+                if let Some(parent) = target.parent() {
+                    std::fs::create_dir_all(parent).map_err(|e| e.to_string())?;
+                }
+                std::fs::write(&target, file.data.as_ref()).map_err(|e| e.to_string())?;
+            }
             std::fs::File::create(&sentinel).map_err(|e| e.to_string())?;
         }
 

From cec3c14aa88ac1296f6578fc0b7c27e2fe01b994 Mon Sep 17 00:00:00 2001
From: patrick <patrick@rtk-ai.app>
Date: Thu, 14 May 2026 22:43:11 +0200
Subject: [PATCH 2/2] fix: import Embed trait for iter/get methods (rust-embed
 v8)

---
 src/backend/piper.rs | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/src/backend/piper.rs b/src/backend/piper.rs
index b031039..10cd5d4 100644
--- a/src/backend/piper.rs
+++ b/src/backend/piper.rs
@@ -8,7 +8,7 @@ use std::sync::{Mutex, OnceLock};
 
 use anyhow::{Context, Result};
 use piper_rs::Piper;
-use rust_embed::RustEmbed;
+use rust_embed::{Embed, RustEmbed};
 
 use super::{SpeakOptions, TtsBackend};
 use crate::config;