Text to subtitles SRT

This is a terminal node - it doesn't feed anything downstream, it writes a .srt subtitle file to disk and that's the end of the graph. It's built for exactly the workflow the README describes for streaming/media use cases: run text through an LLM (translation, script generation, whatever), then use this to turn the timed result into a subtitle file you can attach to a video.

The two required inputs tell you what upstream node this expects to be paired with. translated_texts is the actual subtitle line content - the name suggests it's meant to come out of a translation step in your graph, but functionally it's just the text strings for each subtitle entry. timestamps takes a TIMESTAMP type, which is not a plain string or float - it's a custom type this pack (or something feeding it) defines, meaning you need an upstream node that actually produces TIMESTAMP-typed output, generally something doing speech-to-text or audio transcription with timing information, not something you can hand-type into a text box. If you're trying to use this node and don't have anything in your graph that outputs TIMESTAMP, that's the missing piece - check the pack's other audio/transcription nodes or workflow examples for what produces it, since it isn't one of this pack's text or LLM nodes covered here. output_dir_path is straightforward: where the .srt file gets written, defaulting to a path inside the pack's own output folder - worth changing to something more sensible for your actual project rather than leaving subtitle files buried in custom_nodes/comfyui_LLM_party/output.

This node is is_output_node, meaning ComfyUI treats it as an endpoint of the graph rather than something whose output you'd chain further - there are no outputs listed because none exist. Its whole job is the side effect of writing a file.

Installing it is the same as installing the whole pack: search "comfyui_LLM_party" in ComfyUI Manager, or git clone https://github.com/heshengtao/comfyui_LLM_party into custom_nodes and restart, then pip install -r requirements.txt from inside the pack folder using ComfyUI's own Python (portable builds: python_embeded\python.exe -m pip install -r requirements.txt explicitly, not your system pip).

Where people get stuck: the TIMESTAMP type mismatch is the obvious one - you can't just hand this node a list of numbers or a string with timecodes in it, it wants the actual typed output from whatever upstream node produces it, so if the wire refuses to connect in the ComfyUI editor, that's ComfyUI correctly telling you the types don't match rather than a bug. Second, check output_dir_path actually exists and is writable before you run the graph - a nonexistent directory is a common silent failure point for file-writing nodes in general. And if your subtitles come out misaligned with the audio/video, the fix is almost never in this node - it just writes what it's given, so go back to whatever produced the timestamps and translated_texts inputs and verify those line up correctly first.