feat: add Xiaomi MiMo speech support

README.md

@@ -321,6 +321,7 @@ codewhale --provider openrouter --model minimax/minimax-m3
 # Xiaomi MiMo
 codewhale auth set --provider xiaomi-mimo --api-key "YOUR_XIAOMI_KEY"
 codewhale --provider xiaomi-mimo --model mimo-v2.5-pro
+codewhale --provider xiaomi-mimo speech "Hello from MiMo" --model tts -o hello.wav
 
 # Novita
 codewhale auth set --provider novita --api-key "YOUR_NOVITA_API_KEY"

README.zh-CN.md

@@ -268,6 +268,7 @@ codewhale --provider openrouter --model qwen/qwen3.7-max
 # Xiaomi MiMo
 codewhale auth set --provider xiaomi-mimo --api-key "YOUR_XIAOMI_MIMO_API_KEY"
 codewhale --provider xiaomi-mimo --model mimo-v2.5-pro
+codewhale --provider xiaomi-mimo speech "???MiMo" --model tts -o hello.wav
 
 # Novita
 codewhale auth set --provider novita --api-key "YOUR_NOVITA_API_KEY"

config.example.toml

@@ -45,6 +45,9 @@ base_url = "https://api.deepseek.com/beta"
 #   deepseek-ai/deepseek-v4-flash   — default AtlasCloud model ID
 #   deepseek-reasoner               — default Wanjie Ark model ID
 #   mimo-v2.5-pro                   — default Xiaomi MiMo model ID
+#   mimo-v2.5-tts                   ? Xiaomi MiMo speech/TTS model ID
+#   mimo-v2.5-tts-voicedesign       ? Xiaomi MiMo voice-design TTS model ID
+#   mimo-v2.5-tts-voiceclone        ? Xiaomi MiMo voice-clone TTS model ID
 #   accounts/fireworks/models/deepseek-v4-pro — Fireworks AI Pro model ID
 #   deepseek-ai/DeepSeek-V4-Pro    — SiliconFlow hosted Pro model ID
 #   deepseek-ai/DeepSeek-V4-Flash  — SiliconFlow hosted Flash model ID
@@ -120,6 +123,11 @@ memory_path = "~/.codewhale/memory.md"
 # Parsed but currently unused (reserved for future versions):
 # tools_file = "./tools.json"
 
+# Xiaomi MiMo speech/TTS defaults. Also configurable with
+# XIAOMI_MIMO_SPEECH_OUTPUT_DIR / MIMO_SPEECH_OUTPUT_DIR.
+[speech]
+# output_dir = "./speech"
+
 # Native tool catalog controls (#2076). By default only the core tool surface
 # is loaded into the model context; less common native tools are discoverable
 # through ToolSearch and loaded on first use.
@@ -286,7 +294,9 @@ max_subagents = 10 # optional (1-20)
 [providers.xiaomi_mimo]
 # api_key = "YOUR_XIAOMI_KEY"
 # base_url = "https://api.xiaomimimo.com/v1"
-# model = "mimo-v2.5-pro"
+# model = "mimo-v2.5-pro"              # chat/reasoning
+# TTS aliases are also accepted by `codewhale speech`: tts, voice-design, voice-clone
+# TTS model IDs: mimo-v2.5-tts, mimo-v2.5-tts-voicedesign, mimo-v2.5-tts-voiceclone, mimo-v2-tts
 
 # Novita AI-hosted inference (https://novita.ai)
 [providers.novita]

crates/agent/src/lib.rs

@@ -307,6 +307,46 @@ impl Default for ModelRegistry {
                 supports_tools: true,
                 supports_reasoning: true,
             },
+            ModelInfo {
+                id: "mimo-v2.5-tts".to_string(),
+                provider: ProviderKind::XiaomiMimo,
+                aliases: vec![
+                    "tts".to_string(),
+                    "speech".to_string(),
+                    "mimo-tts".to_string(),
+                ],
+                supports_tools: false,
+                supports_reasoning: false,
+            },
+            ModelInfo {
+                id: "mimo-v2.5-tts-voicedesign".to_string(),
+                provider: ProviderKind::XiaomiMimo,
+                aliases: vec![
+                    "voicedesign".to_string(),
+                    "voice-design".to_string(),
+                    "mimo-voice-design".to_string(),
+                ],
+                supports_tools: false,
+                supports_reasoning: false,
+            },
+            ModelInfo {
+                id: "mimo-v2.5-tts-voiceclone".to_string(),
+                provider: ProviderKind::XiaomiMimo,
+                aliases: vec![
+                    "voiceclone".to_string(),
+                    "voice-clone".to_string(),
+                    "mimo-voice-clone".to_string(),
+                ],
+                supports_tools: false,
+                supports_reasoning: false,
+            },
+            ModelInfo {
+                id: "mimo-v2-tts".to_string(),
+                provider: ProviderKind::XiaomiMimo,
+                aliases: vec!["mimo-v2-speech".to_string()],
+                supports_tools: false,
+                supports_reasoning: false,
+            },
             ModelInfo {
                 id: "deepseek/deepseek-v4-pro".to_string(),
                 provider: ProviderKind::Novita,
@@ -649,6 +689,22 @@ mod tests {
         assert!(resolved.resolved.supports_reasoning);
     }
 
+    #[test]
+    fn xiaomi_mimo_tts_aliases_resolve_when_provider_hinted() {
+        let registry = ModelRegistry::default();
+        let resolved = registry.resolve(Some("tts"), Some(ProviderKind::XiaomiMimo));
+        assert_eq!(resolved.resolved.provider, ProviderKind::XiaomiMimo);
+        assert_eq!(resolved.resolved.id, "mimo-v2.5-tts");
+        assert!(!resolved.resolved.supports_tools);
+        assert!(!resolved.resolved.supports_reasoning);
+
+        let resolved = registry.resolve(Some("voice-design"), Some(ProviderKind::XiaomiMimo));
+        assert_eq!(resolved.resolved.id, "mimo-v2.5-tts-voicedesign");
+
+        let resolved = registry.resolve(Some("voiceclone"), Some(ProviderKind::XiaomiMimo));
+        assert_eq!(resolved.resolved.id, "mimo-v2.5-tts-voiceclone");
+    }
+
     #[test]
     fn wanjie_ark_default_uses_reasoner_model_id() {
         let registry = ModelRegistry::default();

crates/cli/src/lib.rs

@@ -133,6 +133,9 @@ enum Commands {
     Doctor(TuiPassthroughArgs),
     /// List live DeepSeek API models via the TUI binary.
     Models(TuiPassthroughArgs),
+    /// Generate speech audio with Xiaomi MiMo TTS models via the TUI binary.
+    #[command(visible_alias = "tts")]
+    Speech(TuiPassthroughArgs),
     /// List saved TUI sessions.
     Sessions(TuiPassthroughArgs),
     /// Resume a saved TUI session.
@@ -510,6 +513,10 @@ fn run() -> Result<()> {
             let resolved_runtime = resolve_runtime_for_dispatch(&mut store, &runtime_overrides);
             delegate_to_tui(&cli, &resolved_runtime, tui_args("models", args))
         }
+        Some(Commands::Speech(args)) => {
+            let resolved_runtime = resolve_runtime_for_dispatch(&mut store, &runtime_overrides);
+            delegate_to_tui(&cli, &resolved_runtime, tui_args("speech", args))
+        }
         Some(Commands::Sessions(args)) => {
             let resolved_runtime = resolve_runtime_for_dispatch(&mut store, &runtime_overrides);
             delegate_to_tui(&cli, &resolved_runtime, tui_args("sessions", args))

crates/config/src/lib.rs

@@ -42,6 +42,10 @@ const OPENROUTER_TENCENT_HY3_PREVIEW_MODEL: &str = "tencent/hy3-preview";
 const OPENROUTER_XIAOMI_MIMO_V2_5_PRO_MODEL: &str = "xiaomi/mimo-v2.5-pro";
 const OPENROUTER_XIAOMI_MIMO_V2_5_MODEL: &str = "xiaomi/mimo-v2.5";
 const DEFAULT_XIAOMI_MIMO_MODEL: &str = "mimo-v2.5-pro";
+const XIAOMI_MIMO_TTS_MODEL: &str = "mimo-v2.5-tts";
+const XIAOMI_MIMO_TTS_VOICE_DESIGN_MODEL: &str = "mimo-v2.5-tts-voicedesign";
+const XIAOMI_MIMO_TTS_VOICE_CLONE_MODEL: &str = "mimo-v2.5-tts-voiceclone";
+const XIAOMI_MIMO_V2_TTS_MODEL: &str = "mimo-v2-tts";
 const DEFAULT_NOVITA_MODEL: &str = "deepseek/deepseek-v4-pro";
 const DEFAULT_NOVITA_FLASH_MODEL: &str = "deepseek/deepseek-v4-flash";
 const DEFAULT_FIREWORKS_MODEL: &str = "accounts/fireworks/models/deepseek-v4-pro";
@@ -1426,6 +1430,12 @@ pub fn load_project_config(workspace: &Path) -> Option<ConfigToml> {
 }
 
 fn normalize_model_for_provider(provider: ProviderKind, model: &str) -> String {
+    if matches!(provider, ProviderKind::XiaomiMimo)
+        && let Some(canonical) = canonical_xiaomi_mimo_model_id(model)
+    {
+        return canonical.to_string();
+    }
+
     if matches!(
         provider,
         ProviderKind::Atlascloud
@@ -1500,6 +1510,38 @@ fn normalize_model_for_provider(provider: ProviderKind, model: &str) -> String {
     }
 }
 
+fn canonical_xiaomi_mimo_model_id(model: &str) -> Option<&'static str> {
+    let normalized = model.trim().to_ascii_lowercase();
+    let normalized = normalized.replace(['_', ' '], "-");
+    match normalized.as_str() {
+        "mimo"
+        | DEFAULT_XIAOMI_MIMO_MODEL
+        | "mimo-v2-5-pro"
+        | "xiaomi-mimo-v2.5-pro"
+        | "xiaomi-mimo-v2-5-pro" => Some(DEFAULT_XIAOMI_MIMO_MODEL),
+        "mimo-v2.5" | "mimo-v25" | "mimo-v2-5" | "xiaomi-mimo-v2.5" | "xiaomi-mimo-v2-5" => {
+            Some("mimo-v2.5")
+        }
+        "mimo-tts" | "mimo-v25-tts" | "mimo-v2.5-tts" | "tts" | "speech" => {
+            Some(XIAOMI_MIMO_TTS_MODEL)
+        }
+        "mimo-tts-voicedesign"
+        | "mimo-voice-design"
+        | "mimo-v25-tts-voicedesign"
+        | "mimo-v2.5-tts-voicedesign"
+        | "voicedesign"
+        | "voice-design" => Some(XIAOMI_MIMO_TTS_VOICE_DESIGN_MODEL),
+        "mimo-tts-voiceclone"
+        | "mimo-voice-clone"
+        | "mimo-v25-tts-voiceclone"
+        | "mimo-v2.5-tts-voiceclone"
+        | "voiceclone"
+        | "voice-clone" => Some(XIAOMI_MIMO_TTS_VOICE_CLONE_MODEL),
+        "mimo-v2-tts" => Some(XIAOMI_MIMO_V2_TTS_MODEL),
+        _ => None,
+    }
+}
+
 fn canonical_openrouter_recent_model_id(model: &str) -> Option<&'static str> {
     let normalized = model.trim().to_ascii_lowercase();
     let normalized = normalized.replace(['_', ' '], "-");
@@ -3263,6 +3305,26 @@ unix_socket_path = "/tmp/cw-hooks.sock"
         assert_eq!(resolved.model, DEFAULT_XIAOMI_MIMO_MODEL);
     }
 
+    #[test]
+    fn xiaomi_mimo_tts_aliases_resolve_to_canonical_models() {
+        assert_eq!(
+            normalize_model_for_provider(ProviderKind::XiaomiMimo, "tts"),
+            "mimo-v2.5-tts"
+        );
+        assert_eq!(
+            normalize_model_for_provider(ProviderKind::XiaomiMimo, "voice-design"),
+            "mimo-v2.5-tts-voicedesign"
+        );
+        assert_eq!(
+            normalize_model_for_provider(ProviderKind::XiaomiMimo, "voiceclone"),
+            "mimo-v2.5-tts-voiceclone"
+        );
+        assert_eq!(
+            normalize_model_for_provider(ProviderKind::XiaomiMimo, "custom-mimo-model"),
+            "custom-mimo-model"
+        );
+    }
+
     #[test]
     fn novita_provider_defaults_to_canonical_endpoint_and_model() {
         let _lock = env_lock();