What happened?
I got "MiniMax H3 screenplay dialogue needs 159 spoken words, but the selected windows safely fit 30. Increase total duration or add windows before generating."
Your system (OS / GPU / RAM / VRAM / etc.)
Windows 11 128GbGB DDR4 RTX 4070 12GB VRAM
Logs / full error output
[LLM] Connected to remote provider: model=qwen/qwen3.5-9b, url=http://127.0.0.1:1234
[LLM] json_schema requested but provider=remote — sending unconstrained (grammar is local llama-server only)
[LLM] Dropped 3 llama.cpp-only field(s) not accepted by provider=remote: cache_prompt, chat_template_kwargs, enable_thinking
[LLM] Response: 1000 tokens generated (prompt=1526, finish=length)
[LLM] WARNING: Server returned empty content despite generating 1000 tokens (model likely consumed all tokens on internal reasoning)
[LLM] Reasoning content detected (4513 chars) — model used thinking mode.
[LLM] Completion telemetry: reasoning=1000 tokens, answer=0 tokens, effort=off/default, budget=0, finish=length
[MiniMax H3] Shared-treatment fallback: screenplay dialogue uses 159 words; the selected windows safely fit 30; segment 1 dialogue uses 124 words; budget is 19; segment 2 dialogue uses 35 words; budget is 11
Traceback (most recent call last):
File "C:\pinokio\api\Maestro.git\app\launch.py", line 8351, in llm_plan_h3_sequence
result = await asyncio.to_thread(
^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERD030\AppData\Roaming\uv\python\cpython-3.11.14-windows-x86_64-none\Lib\asyncio\threads.py", line 25, in to_thread
return await loop.run_in_executor(None, func_call)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERD030\AppData\Roaming\uv\python\cpython-3.11.14-windows-x86_64-none\Lib\concurrent\futures\thread.py", line 58, in run
result = self.fn(*self.args, **self.kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_sequence_planner.py", line 1627, in plan_h3_reference_sequence
staged = plan_h3_story_segments(
^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_story_ledger.py", line 6410, in plan_h3_story_segments
) = _prepare_render_dialogue_schedule(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_story_ledger.py", line 2329, in _prepare_render_dialogue_schedule
raise H3DialogueTimingError(
services.h3_story_ledger.H3DialogueTimingError: MiniMax H3 screenplay dialogue needs 159 spoken words, but the selected windows safely fit 30. Increase total duration or add windows before generating.
What happened?
I got "MiniMax H3 screenplay dialogue needs 159 spoken words, but the selected windows safely fit 30. Increase total duration or add windows before generating."
Your system (OS / GPU / RAM / VRAM / etc.)
Windows 11 128GbGB DDR4 RTX 4070 12GB VRAM
Logs / full error output
[LLM] Connected to remote provider: model=qwen/qwen3.5-9b, url=http://127.0.0.1:1234
[LLM] json_schema requested but provider=remote — sending unconstrained (grammar is local llama-server only)
[LLM] Dropped 3 llama.cpp-only field(s) not accepted by provider=remote: cache_prompt, chat_template_kwargs, enable_thinking
[LLM] Response: 1000 tokens generated (prompt=1526, finish=length)
[LLM] WARNING: Server returned empty content despite generating 1000 tokens (model likely consumed all tokens on internal reasoning)
[LLM] Reasoning content detected (4513 chars) — model used thinking mode.
[LLM] Completion telemetry: reasoning=1000 tokens, answer=0 tokens, effort=off/default, budget=0, finish=length
[MiniMax H3] Shared-treatment fallback: screenplay dialogue uses 159 words; the selected windows safely fit 30; segment 1 dialogue uses 124 words; budget is 19; segment 2 dialogue uses 35 words; budget is 11
Traceback (most recent call last):
File "C:\pinokio\api\Maestro.git\app\launch.py", line 8351, in llm_plan_h3_sequence
result = await asyncio.to_thread(
^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERD030\AppData\Roaming\uv\python\cpython-3.11.14-windows-x86_64-none\Lib\asyncio\threads.py", line 25, in to_thread
return await loop.run_in_executor(None, func_call)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\Users\USERD030\AppData\Roaming\uv\python\cpython-3.11.14-windows-x86_64-none\Lib\concurrent\futures\thread.py", line 58, in run
result = self.fn(*self.args, **self.kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_sequence_planner.py", line 1627, in plan_h3_reference_sequence
staged = plan_h3_story_segments(
^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_story_ledger.py", line 6410, in plan_h3_story_segments
) = _prepare_render_dialogue_schedule(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "C:\pinokio\api\Maestro.git\app\services\h3_story_ledger.py", line 2329, in _prepare_render_dialogue_schedule
raise H3DialogueTimingError(
services.h3_story_ledger.H3DialogueTimingError: MiniMax H3 screenplay dialogue needs 159 spoken words, but the selected windows safely fit 30. Increase total duration or add windows before generating.