agent/codex
1
0
Fork 0
mirror of https://github.com/pchuan98/codex.git synced 2026-07-01 00:31:56 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add realtime audio device config (#12849)

Browse Source 
## Summary
- add top-level realtime audio config for microphone and speaker
selection
- apply configured devices when starting realtime capture and playback
- keep missing-device behavior on the system default fallback path

## Validation
- just write-config-schema
- cargo test -p codex-core realtime_audio
- cargo test -p codex-tui
- just fix -p codex-core
- just fix -p codex-tui
- just fmt

---------

Co-authored-by: Codex <noreply@openai.com>
This commit is contained in:
Ahmed Ibrahim
2026-02-26 15:08:21 -08:00
committed by GitHub
Unverified
parent fd719d3828
commit a0e86c69fe
7 changed files with 299 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files
codex-rs/core/src/config/edit.rs
+68
View File
@@ -839,6 +839,30 @@ impl ConfigEditsBuilder {
self
}
pub fn set_realtime_microphone(mut self, microphone: Option<&str>) -> Self {
let segments = vec!["audio".to_string(), "microphone".to_string()];
match microphone {
Some(microphone) => self.edits.push(ConfigEdit::SetPath {
segments,
value: value(microphone),
}),
None => self.edits.push(ConfigEdit::ClearPath { segments }),
}
self
}
pub fn set_realtime_speaker(mut self, speaker: Option<&str>) -> Self {
let segments = vec!["audio".to_string(), "speaker".to_string()];
match speaker {
Some(speaker) => self.edits.push(ConfigEdit::SetPath {
segments,
value: value(speaker),
}),
None => self.edits.push(ConfigEdit::ClearPath { segments }),
}
self
}
pub fn clear_legacy_windows_sandbox_keys(mut self) -> Self {
for key in [
"experimental_windows_sandbox",
@@ -1804,6 +1828,50 @@ model_reasoning_effort = "high"
assert_eq!(notice, Some(true));
}
#[test]
fn blocking_builder_set_realtime_audio_persists_and_clears() {
let tmp = tempdir().expect("tmpdir");
let codex_home = tmp.path();
ConfigEditsBuilder::new(codex_home)
.set_realtime_microphone(Some("USB Mic"))
.set_realtime_speaker(Some("Desk Speakers"))
.apply_blocking()
.expect("persist realtime audio");
let raw = std::fs::read_to_string(codex_home.join(CONFIG_TOML_FILE)).expect("read config");
let config: TomlValue = toml::from_str(&raw).expect("parse config");
let realtime_audio = config
.get("audio")
.and_then(TomlValue::as_table)
.expect("audio table should exist");
assert_eq!(
realtime_audio.get("microphone").and_then(TomlValue::as_str),
Some("USB Mic")
);
assert_eq!(
realtime_audio.get("speaker").and_then(TomlValue::as_str),
Some("Desk Speakers")
);
ConfigEditsBuilder::new(codex_home)
.set_realtime_microphone(None)
.apply_blocking()
.expect("clear realtime microphone");
let raw = std::fs::read_to_string(codex_home.join(CONFIG_TOML_FILE)).expect("read config");
let config: TomlValue = toml::from_str(&raw).expect("parse config");
let realtime_audio = config
.get("audio")
.and_then(TomlValue::as_table)
.expect("audio table should exist");
assert_eq!(realtime_audio.get("microphone"), None);
assert_eq!(
realtime_audio.get("speaker").and_then(TomlValue::as_str),
Some("Desk Speakers")
);
}
#[test]
fn replace_mcp_servers_blocking_clears_table_when_empty() {
let tmp = tempdir().expect("tmpdir");
codex-rs/core/src/config/mod.rs
+63
View File
@@ -426,6 +426,9 @@ pub struct Config {
/// Base URL for requests to ChatGPT (as opposed to the OpenAI API).
pub chatgpt_base_url: String,
/// Machine-local realtime audio device preferences used by realtime voice.
pub realtime_audio: RealtimeAudioConfig,
/// Experimental / do not use. Overrides only the realtime conversation
/// websocket transport base URL (the `Op::RealtimeConversation` `/ws`
/// connection) without changing normal provider HTTP requests.
@@ -1175,6 +1178,10 @@ pub struct ConfigToml {
/// Base URL for requests to ChatGPT (as opposed to the OpenAI API).
pub chatgpt_base_url: Option<String>,
/// Machine-local realtime audio device preferences used by realtime voice.
#[serde(default)]
pub audio: Option<RealtimeAudioToml>,
/// Experimental / do not use. Overrides only the realtime conversation
/// websocket transport base URL (the `Op::RealtimeConversation` `/ws`
/// connection) without changing normal provider HTTP requests.
@@ -1306,6 +1313,19 @@ impl ProjectConfig {
}
}
#[derive(Debug, Clone, Default, PartialEq, Eq)]
pub struct RealtimeAudioConfig {
pub microphone: Option<String>,
pub speaker: Option<String>,
}
#[derive(Serialize, Deserialize, Debug, Clone, Default, PartialEq, Eq, JsonSchema)]
#[schemars(deny_unknown_fields)]
pub struct RealtimeAudioToml {
pub microphone: Option<String>,
pub speaker: Option<String>,
}
#[derive(Serialize, Deserialize, Debug, Clone, Default, PartialEq, JsonSchema)]
#[schemars(deny_unknown_fields)]
pub struct ToolsToml {
@@ -2146,6 +2166,12 @@ impl Config {
.chatgpt_base_url
.or(cfg.chatgpt_base_url)
.unwrap_or("https://chatgpt.com/backend-api/".to_string()),
realtime_audio: cfg
.audio
.map_or_else(RealtimeAudioConfig::default, |audio| RealtimeAudioConfig {
microphone: audio.microphone,
speaker: audio.speaker,
}),
experimental_realtime_ws_base_url: cfg.experimental_realtime_ws_base_url,
experimental_realtime_ws_backend_prompt: cfg.experimental_realtime_ws_backend_prompt,
forced_chatgpt_workspace_id,
@@ -4766,6 +4792,7 @@ model_verbosity = "high"
model_verbosity: None,
personality: Some(Personality::Pragmatic),
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
realtime_audio: RealtimeAudioConfig::default(),
experimental_realtime_ws_base_url: None,
experimental_realtime_ws_backend_prompt: None,
base_instructions: None,
@@ -4892,6 +4919,7 @@ model_verbosity = "high"
model_verbosity: None,
personality: Some(Personality::Pragmatic),
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
realtime_audio: RealtimeAudioConfig::default(),
experimental_realtime_ws_base_url: None,
experimental_realtime_ws_backend_prompt: None,
base_instructions: None,
@@ -5016,6 +5044,7 @@ model_verbosity = "high"
model_verbosity: None,
personality: Some(Personality::Pragmatic),
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
realtime_audio: RealtimeAudioConfig::default(),
experimental_realtime_ws_base_url: None,
experimental_realtime_ws_backend_prompt: None,
base_instructions: None,
@@ -5126,6 +5155,7 @@ model_verbosity = "high"
model_verbosity: Some(Verbosity::High),
personality: Some(Personality::Pragmatic),
chatgpt_base_url: "https://chatgpt.com/backend-api/".to_string(),
realtime_audio: RealtimeAudioConfig::default(),
experimental_realtime_ws_base_url: None,
experimental_realtime_ws_backend_prompt: None,
base_instructions: None,
@@ -5970,6 +6000,39 @@ experimental_realtime_ws_backend_prompt = "prompt from config"
);
Ok(())
}
#[test]
fn realtime_audio_loads_from_config_toml() -> std::io::Result<()> {
let cfg: ConfigToml = toml::from_str(
r#"
[audio]
microphone = "USB Mic"
speaker = "Desk Speakers"
"#,
)
.expect("TOML deserialization should succeed");
let realtime_audio = cfg
.audio
.as_ref()
.expect("realtime audio config should be present");
assert_eq!(realtime_audio.microphone.as_deref(), Some("USB Mic"));
assert_eq!(realtime_audio.speaker.as_deref(), Some("Desk Speakers"));
let codex_home = TempDir::new()?;
let config = Config::load_from_base_config_with_overrides(
cfg,
ConfigOverrides::default(),
codex_home.path().to_path_buf(),
)?;
assert_eq!(config.realtime_audio.microphone.as_deref(), Some("USB Mic"));
assert_eq!(
config.realtime_audio.speaker.as_deref(),
Some("Desk Speakers")
);
Ok(())
}
}
#[cfg(test)]