feat: add support for custom OpenAI-compatible endpoints in config UI

app.py

@@ -66,6 +66,7 @@ _USER_CONFIG_LOCK = threading.Lock()
 _USER_CONFIG: Dict[str, Optional[str]] = {
     "hf_token": None,
     "model": None,
+    "custom_endpoint": None,
 }
 
 
@@ -79,6 +80,11 @@ def get_user_model() -> Optional[str]:
         return _USER_CONFIG["model"]
 
 
+def get_user_custom_endpoint() -> Optional[str]:
+    with _USER_CONFIG_LOCK:
+        return _USER_CONFIG["custom_endpoint"]
+
+
 # ---------------------------------------------------------------------------
 # Llama.cpp Inference Setup
 # ---------------------------------------------------------------------------
@@ -252,6 +258,7 @@ def generate_llm_story_beat(genre: str, history: List[Dict[str, str]], instructi
             temperature=0.7,
             token=get_user_hf_token(),
             model=get_user_model(),
+            custom_endpoint=get_user_custom_endpoint(),
         )
         return result.text.strip()
     except Exception as e:
@@ -276,6 +283,7 @@ def generate_llm_choices_for_story(genre: str, story: str) -> List[str]:
             temperature=0.8,
             token=get_user_hf_token(),
             model=get_user_model(),
+            custom_endpoint=get_user_custom_endpoint(),
         )
         return _parse_choices(result.text)
     except Exception as e:
@@ -307,6 +315,7 @@ def generate_llm_health_effect(genre: str, history: List[Dict[str, str]], curren
             temperature=0.8,
             token=get_user_hf_token(),
             model=get_user_model(),
+            custom_endpoint=get_user_custom_endpoint(),
         )
         text = result.text.strip()
     except Exception as e:
@@ -559,15 +568,18 @@ async def game_choice(payload: dict):
 @fastapi_app.post("/api/config")
 async def update_config(body: dict):
     with _USER_CONFIG_LOCK:
-        if body.get("hf_token"):
-            _USER_CONFIG["hf_token"] = body["hf_token"].strip() or None
-        if body.get("model") and str(body["model"]).strip():
-            _USER_CONFIG["model"] = str(body["model"]).strip()
+        if "hf_token" in body:
+            _USER_CONFIG["hf_token"] = body["hf_token"].strip() if body["hf_token"] else None
+        if "model" in body:
+            _USER_CONFIG["model"] = body["model"].strip() if body["model"] else None
+        if "custom_endpoint" in body:
+            _USER_CONFIG["custom_endpoint"] = body["custom_endpoint"].strip() if body["custom_endpoint"] else None
         current = dict(_USER_CONFIG)
     return {
         "status": "ok",
         "model": current["model"] or TINYBARD_MODEL,
         "has_token": bool(current["hf_token"]),
+        "custom_endpoint": current["custom_endpoint"],
     }
 
 
@@ -578,6 +590,7 @@ async def get_config():
     return {
         "model": current["model"] or TINYBARD_MODEL,
         "has_token": bool(current["hf_token"]),
+        "custom_endpoint": current["custom_endpoint"],
     }
 
 

shared/inference_client.py

@@ -111,6 +111,7 @@ def generate(
     temperature: float = 0.7,
     token: Optional[str] = None,
     model: Optional[str] = None,
+    custom_endpoint: Optional[str] = None,
 ) -> InferenceResult:
     """Run a chat-style inference call, with cooldown enforcement.
 

static/index.html

@@ -85,6 +85,9 @@
             <label for="tb-token-input">HF Token (optional)</label>
             <input type="password" id="tb-token-input" placeholder="hf_..." />
 
+            <label for="tb-endpoint-input">Custom Endpoint (optional)</label>
+            <input type="text" id="tb-endpoint-input" placeholder="https://api.together.xyz/v1" />
+
             <button id="tb-config-save">Save & Reload</button>
             <span id="tb-config-status" style="margin-left:0.5rem;color:var(--asp-moss,#588157);"></span>
         </div>

static/main.js

@@ -253,6 +253,7 @@ const configClose = document.getElementById('tb-config-close');
 const configSave = document.getElementById('tb-config-save');
 const modelInput = document.getElementById('tb-model-input');
 const tokenInput = document.getElementById('tb-token-input');
+const endpointInput = document.getElementById('tb-endpoint-input');
 const configStatus = document.getElementById('tb-config-status');
 
 if (configBtn && configModal) {
@@ -260,6 +261,9 @@ if (configBtn && configModal) {
         const cfg = await fetch('/api/config').then(r => r.json());
         modelInput.value = cfg.model || '';
         tokenInput.value = '';
+        if (endpointInput) {
+            endpointInput.value = cfg.custom_endpoint || '';
+        }
         configStatus.textContent = '';
         configModal.style.display = 'flex';
     });
@@ -274,8 +278,12 @@ if (configBtn && configModal) {
 
     configSave.addEventListener('click', async () => {
         const body = {};
-        if (modelInput.value.trim()) body.model = modelInput.value.trim();
-        if (tokenInput.value.trim()) body.hf_token = tokenInput.value.trim();
+        // We always pass these fields to let user clear them (by passing empty strings or letting backend handle them)
+        body.model = modelInput.value.trim() || "";
+        body.hf_token = tokenInput.value.trim() || "";
+        if (endpointInput) {
+            body.custom_endpoint = endpointInput.value.trim() || "";
+        }
 
         const resp = await fetch('/api/config', {
             method: 'POST',