diff --git a/llmage/llmclient.py b/llmage/llmclient.py
index 5f95ea0..d322b0c 100644
--- a/llmage/llmclient.py
+++ b/llmage/llmclient.py
@@ -388,7 +388,8 @@ async def inference_generator(request, *args, params_kw=None, **kw):
 		yield d
 
 async def _inference_generator(request, callerid, callerorgid, 
-							params_kw=params_kw, **kw):
+							params_kw={}, **kw):
+	env = request._run_ns
 	if not params_kw:
 		params_kw = env.params_kw
 	if not params_kw.transno:
diff --git a/wwwroot/llminference.dspy b/wwwroot/llminference.dspy
index 4baf6a9..e1b20fa 100644
--- a/wwwroot/llminference.dspy
+++ b/wwwroot/llminference.dspy
@@ -1,7 +1,9 @@
 debug(f'{params_kw=}')
 userid = await get_user()
+userorgid = await get_userorgid()
 if userid is None:
 	return UiError(title='llm inference', message='Please login first')
+f = await checkCustomerBalance(params_kw.llmid, userorgid)
 kdbids = params_kw.kdbids
 if kdbids:
 	data = {