Add HF token input for ZeroGPU Pro quota on regen

Users can paste their HF token in the Settings accordion so that
segment regen calls use Authorization: Bearer <token> — the documented
way for ZeroGPU to attribute quota to a Pro account. Falls back to
x-ip-token relay if no token is provided. Quota now correctly debits
the logged-in user's Pro pool instead of the shared anonymous pool.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

app.py

@@ -2133,13 +2133,26 @@ _GLOBAL_JS = """
       console.warn('[fireRegen] fn_index not found for api_name:', apiName);
       return;
     }
-    // Fetch a fresh x-ip-token immediately before queuing (JWT expires ~170s,
-    // so we always grab a new one to ensure Pro quota attribution).
-    _fetchIpToken().then(function(ipToken) {
+    // Build auth headers for the regen call.
+    // Prefer a user-supplied HF token (Authorization: Bearer) which is the
+    // documented way for ZeroGPU to attribute quota to a logged-in Pro account.
+    // Fall back to the x-ip-token relay approach if no HF token is provided.
+    var userHfToken = '';
+    var hfTokenEl = document.getElementById('hf_token_input');
+    if (hfTokenEl) { var inp = hfTokenEl.querySelector('input,textarea'); if (inp) userHfToken = (inp.value || '').trim(); }
+
+    var _doRegen = function(ipToken) {
       var regenHeaders = {'Content-Type': 'application/json'};
-      if (ipToken) { regenHeaders['x-ip-token'] = ipToken; console.log('[fireRegen] using fresh x-ip-token, len:', ipToken.length); }
-      else { console.warn('[fireRegen] no x-ip-token available, regen may use anonymous quota'); }
-      return fetch('/gradio_api/queue/join', {
+      if (userHfToken) {
+        regenHeaders['Authorization'] = 'Bearer ' + userHfToken;
+        console.log('[fireRegen] using HF token for Pro quota attribution');
+      } else if (ipToken) {
+        regenHeaders['x-ip-token'] = ipToken;
+        console.log('[fireRegen] using fresh x-ip-token, len:', ipToken.length);
+      } else {
+        console.warn('[fireRegen] no auth available, regen may use anonymous quota');
+      }
+      fetch('/gradio_api/queue/join', {
         method: 'POST',
         credentials: 'include',
         headers: regenHeaders,
@@ -2150,18 +2163,23 @@ _GLOBAL_JS = """
           event_data: null,
           trigger_id: null
         })
+      }).then(function(r) { return r.json(); }).then(function(j) {
+        if (!j.event_id) { console.error('[fireRegen] no event_id:', j); return; }
+        console.log('[fireRegen] queued, event_id:', j.event_id);
+        _listenAndApply(j.event_id, slot_id, seg_idx, _preRegenWaveHtml, _preRegenVideoSrc);
+      }).catch(function(e) {
+        console.error('[fireRegen] fetch error:', e);
+        if (lbl) lbl.textContent = 'Error — see console';
+        var sb = document.getElementById('wf_statusbar_' + slot_id);
+        if (sb) { sb.style.color = '#e05252'; sb.textContent = '\u26a0 Request failed: ' + e.message; }
       });
-    }).then(function(r) { return r.json(); }).then(function(j) {
-      if (!j.event_id) { console.error('[fireRegen] no event_id:', j); return; }
-      console.log('[fireRegen] queued, event_id:', j.event_id);
-      // Subscribe to SSE stream and apply outputs when ready
-      _listenAndApply(j.event_id, slot_id, seg_idx, _preRegenWaveHtml, _preRegenVideoSrc);
-    }).catch(function(e) {
-      console.error('[fireRegen] fetch error:', e);
-      if (lbl) lbl.textContent = 'Error — see console';
-      var sb = document.getElementById('wf_statusbar_' + slot_id);
-      if (sb) { sb.style.color = '#e05252'; sb.textContent = '\u26a0 Request failed: ' + e.message; }
-    });
+    };
+    // If user provided HF token, skip x-ip-token relay (token is sufficient)
+    if (userHfToken) {
+      _doRegen('');
+    } else {
+      _fetchIpToken().then(_doRegen);
+    }
   }
 
   // Subscribe to Gradio SSE stream for an event and apply outputs to DOM.
@@ -2341,6 +2359,19 @@ _GLOBAL_JS = """
 """
 
 with gr.Blocks(title="Generate Audio for Video", css=_SLOT_CSS, js=_GLOBAL_JS) as demo:
+    with gr.Accordion("⚙️ Settings (optional)", open=False):
+        gr.Markdown(
+            "**HF Token** — Paste your [Hugging Face token](https://huggingface.co/settings/tokens) here so that "
+            "segment regeneration uses your own ZeroGPU quota (PRO = 25 min/day) instead of the shared free pool. "
+            "The token is stored only in your browser and never sent anywhere except the Space's own API."
+        )
+        hf_token_input = gr.Textbox(
+            label="Hugging Face Token (optional)",
+            placeholder="hf_...",
+            type="password",
+            elem_id="hf_token_input",
+            max_lines=1,
+        )
     gr.Markdown(
         "# Generate Audio for Video\n"
         "Choose a model and upload a video to generate synchronized audio.\n\n"