Give TTS synthesis its own (longer) request timeout

Long insights are chunked + synthesized server-side and can run past the shared
180s chat/embedding client timeout, causing spurious timeouts. /tts/speech now
uses a per-request timeout from LLAMA_SWAP_TTS_REQUEST_TIMEOUT_SECONDS
(default 600), overriding the client default without affecting chat/embeddings.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

src/ai/llamacpp.rs

@@ -170,9 +170,19 @@ impl LlamaCppClient {
             body["temperature"] = json!(x);
         }
 
+        // TTS gets its own (longer) timeout: synthesizing a long, internally
+        // chunked insight can take minutes, well past the shared chat/embedding
+        // client timeout. Per-request `.timeout()` overrides the client default.
+        let tts_timeout = std::env::var("LLAMA_SWAP_TTS_REQUEST_TIMEOUT_SECONDS")
+            .ok()
+            .and_then(|v| v.parse::<u64>().ok())
+            .filter(|n| *n > 0)
+            .unwrap_or(600);
+
         let resp = self
             .client
             .post(&url)
+            .timeout(Duration::from_secs(tts_timeout))
             .json(&body)
             .send()
             .await