wan / v2.5 / fal / image-to-video
The first closed Wan, via fal on Civitai's cloud
- source_image
- loras
- images
- api_config
- video
- workflow_id
- raw_json
Wan 2.5 is the point where Alibaba's Wan went closed - no open weights ever shipped for it. Which makes this node the only way most people will ever touch it: a cloud recipe in the official civitai-comfy-nodes pack that runs Wan 2.5 image-to-video on fal's fleet through Civitai's Orchestration API. First frame in, native VIDEO out, billed in Buzz.
The honest context: Wan 2.5 was widely read as a modest step up from the open 2.2 - real but not transformative - and the community's mood toward the closed Wans soured as 2.6 and 2.7 underwhelmed. This node's job is to give you 2.5's improvements without asking you to defend Alibaba's open-weight decision. And it gets one thing the older Wan nodes don't: 1080p.
The inputs that matter
prompt and cfg_scale (default 4) are required. Then:
source_image/images- your starting frame. The core of the i2v node.resolution-480p,720p, or1080p(default 1080p). This is the headline upgrade over the fal-hosted 2.1/2.2 nodes, which cap at 720p.aspect_ratio-16:9,9:16, or1:1(default 16:9). Note it's a smaller list than 2.2's fal nodes.enable_prompt_expansion- on by default here (2.2's fal nodes default it off). Falcon-side prompt rewriting is the house style for 2.5; flip it off if you've written exact language.negative_prompt- steering.loras-CIVITAI_LORASsocket; trigger words go inprompt.duration1–30s (default 5),frame_rate24,steps10–50,seed,api_config.
Notice what's not here: no shift, no use_turbo, no interpolator_model, no enable_safety_checker. 2.5's fal tier is a simpler interface than 2.2's - you get resolution, ratio, prompt expansion, and a negative prompt. If you want the dials, that's the 2.2 nodes.
Outputs
Standard pack trio: video (VIDEO) into a Save Video / preview node, workflow_id (STRING) for the Civitai sidebar, raw_json (STRING) for response and cost.
Install and auth
One-time, pack-wide. ComfyUI Manager → search Civitai Comfy Nodes → install, restart:
cd ComfyUI/custom_nodes
git clone https://github.com/civitai/civitai-comfy-nodes.git
pip install -r civitai-comfy-nodes/requirements.txt # just `requests`
Auth: Civitai Auth node, CIVITAI_API_TOKEN (headless), or sidebar sign-in.
Troubleshooting and takes
- 1080p is the reason to pick this over 2.2. If you don't need it, Wan 2.2 via fal is open-weights, cheaper, and at least as well-loved. 2.5's edge is resolution and a small quality bump, not a new league.
- Prompt expansion defaults on. If output ignores your exact wording, that's the feature; turn
enable_prompt_expansionoff. - Closed model, closed filter. No local abliteration exists for 2.5 - Alibaba's moderation applies server-side. A refusal is the model, not the node.
- Slow / timeout. Pack launched June 2026 during Civitai's GPU crunch; long queues and timeouts were the top complaints, and 1080p makes it worse. Default timeout 30 min - raise via Civitai Auth
timeout_minutesorCIVITAI_COMFY_TIMEOUT. - Early preview. Generated nodes "may change without notice."
Reach for this when you specifically want Wan 2.5 at 1080p on a first frame and have accepted you're paying per clip for a model you can't download. The text-to-video sibling is the same recipe without the anchor image.
Inputs (14)
| Name | Type | Default | Description |
|---|---|---|---|
| prompt | STRING | — | |
| cfg_scale | FLOAT | 4.000–100 | — |
| source_imageopt | IMAGE | Either A URL, A DataURL or a Base64 string | |
| frame_rateopt | INT | 240–2147483647 | — |
| durationopt | INT | 51–30 | — |
| seedopt | INT | 00–4294967295 | — |
| stepsopt | INT | 2010–50 | — |
| lorasopt | CIVITAI_LORAS | — | |
| resolutionopt | COMBO | 1080p | 3 options: 480p, 720p, 1080p |
| aspect_ratioopt | COMBO | 16:9 | 3 options: 16:9, 9:16, 1:1 |
| enable_prompt_expansionopt | BOOLEAN | true | — |
| negative_promptopt | STRING | — | |
| imagesopt | IMAGE | — | |
| api_configopt | CIVITAI_CONFIG | Optional Civitai Auth connection; defaults to CIVITAI_API_TOKEN or stored OAuth login. |
Outputs (3)
| Name | Type | Description |
|---|---|---|
| video | VIDEO | — |
| workflow_id | STRING | — |
| raw_json | STRING | — |