FL Time Line

Worth saying plainly up front, because the node's own README entry says it too: FL_TimeLine is a work-in-progress scaffold. It has a full-looking set of controls - IPAdapter preset selection, video dimensions, interpolation mode, frame count and FPS - and even an API endpoint for timeline data behind the scenes. But per the pack's own documentation, it "currently returns the input MODEL unchanged." Every control on this node is wired up in the UI and not yet wired up to actually affect anything downstream.

It lives in Fill-Nodes' "Work-in-Progress" section, alongside a handful of other genuinely unfinished nodes (a fractal recursive sampler, an interactive path animator, a Qwen-Image-Edit strength variant). Fill-Nodes is a big, actively developed pack from one author (filliptm), and this is one of the honest markers of things still being built - better to know that going in than to spend an hour wondering why a keyframe timeline setup isn't changing your output.

How it works

Right now: it accepts a MODEL and passes it straight through, unchanged. The timeline_data string and the other configuration fields exist on the node - presumably feeding the in-progress API endpoint the README mentions, likely intended for some kind of interactive timeline-editing UI similar to how FL_PathAnimator pops a modal editor elsewhere in this pack's WIP section - but none of it currently reaches the model output.

The inputs and outputs that matter

• model - required, and also what comes back out unchanged.
• timeline_data - a multiline string, presumably intended to carry keyframe/timeline structure once the feature is functional.
• Optional: ipadapter_preset (six IPAdapter strength/scope presets, from lightweight SD1.5-only to portrait-focused "PLUS FACE"), video_width/video_height (default 512, up to 2048), interpolation_mode (Linear/Ease_in/Ease_out/Ease_in_out), number_animation_frames (default 96), frames_per_second (default 12) - all configuration for a feature that isn't active yet.

Output is a single MODEL, currently identical to whatever you fed in.

How to install it

ComfyUI Manager: search "Fill-Nodes", install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/filliptm/ComfyUI_Fill-Nodes
pip install -r ComfyUI_Fill-Nodes/requirements.txt

It ships with the rest of the pack; there's no way to install just this one node, and no extra dependency beyond what Fill-Nodes already needs.

Common issues & troubleshooting

"I set up a whole timeline and nothing changed in my output." That's expected right now, not a misconfiguration on your end - the node accepts the settings but doesn't yet apply them. Don't spend time debugging your timeline_data JSON structure or your interpolation mode; the pipe from those settings to actual model behavior isn't connected yet.

Looking for real timeline/keyframe animation control today. This isn't functional yet for that. Depending on what you need, other parts of Fill-Nodes are further along - FL_KsamplerContextWindow for context-window video sampling, or the WAN video-continuation nodes - and are worth checking instead if you need something that works right now.

Wondering if it'll ever finish. No timeline is given in the README beyond the WIP label - this is a case of "keep an eye on the pack's changelog" rather than something this article can promise a date for.