TokiToki Internal
Atlas + S3
All runs
ErrorComfyUI

d9341ca2-a258-457a-ae47-826aa6644040

Queued
Aug 14, 2026, 06:58 PM
Completed
Aug 14, 2026, 06:58 PM
Execution time
764 ms
Outputs
0 · 0 B
Inputs
1
Cached nodes
1

Failure

/home/ec2-user/ComfyUI/custom_nodes/ComfyUI-Grok-Imagine-Video.xai_client.XAIImagineErrorin CuiMaoGrokImageToVideo (node 1)

xAI API returned HTTP 400: {"code": "invalid-argument", "error": "Prompt length exceeds the maximum allowed length of 4096"}

Stack trace
  File "/home/ec2-user/ComfyUI/execution.py", line 542, in execute
    output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/execution.py", line 341, in get_output_data
    return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/execution.py", line 315, in _async_map_node_over_list
    await process_inputs(input_dict, i)
  File "/home/ec2-user/ComfyUI/execution.py", line 303, in process_inputs
    result = f(**inputs)
             ^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/custom_nodes/ComfyUI-Grok-Imagine-Video/nodes.py", line 169, in run
    return _submit_and_download(kwargs, "generation", payload, "grok_image_to_video")
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/custom_nodes/ComfyUI-Grok-Imagine-Video/nodes.py", line 301, in _submit_and_download
    request_id = client.start_video_generation(payload)
                 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/custom_nodes/ComfyUI-Grok-Imagine-Video/xai_client.py", line 88, in start_video_generation
    data = self._request_json("POST", "/v1/videos/generations", json_body=payload)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/ec2-user/ComfyUI/custom_nodes/ComfyUI-Grok-Imagine-Video/xai_client.py", line 171, in _request_json
    raise XAIImagineError(f"xAI API returned HTTP {response.status_code}: {_response_text(response)}")

Outputs

0

The run failed before producing any output.

Text prompts

2
Promptnode 1 · CuiMaoGrokImageToVideo · promptCuiMao Grok Image to Video

Create a NEW photorealistic live-action historical Indian epic sequence using @Video 1 as the PRIMARY CINEMATIC AND MOTION REFERENCE. IMPORTANT: @Video 1 is the direct visual and motion blueprint for the NEW VIDEO. Recreate its camera movement, camera angles, timing, framing, shot progression, character blocking, movement, spatial relationships, background activity, crowd movement, environmental behavior, performance rhythm, cinematic pacing and overall visual language. Do NOT simply copy or replay @Video 1. Create a NEW scene using its cinematic language, adapting the camera, movement and visual progression specifically to @Image 1's dialogue and performance. Characters and faces visible in @Video 1 may remain visible wherever naturally appropriate. However, the MAIN FOCUS must remain on @Image 1 and her complete dialogue. @Image 1 is the MASTER CHARACTER REFERENCE. Preserve her exact facial identity, facial features, hairstyle, costume, jewelry, body proportions, skin tone and overall appearance. Do NOT redesign, reinterpret or alter her. MAIN PERFORMANCE: @Image 1 must perform like a genuine live-action historical-epic actress. Use realistic eye movement, subtle facial expressions, natural breathing, controlled head movement, realistic pauses, accurate lip movement, restrained emotional changes, believable body language and authentic royal authority. Do NOT overact. Do NOT make expressions artificially dramatic. Emotions must develop naturally according to the dialogue. 20-SECOND SEQUENCE: SHOT 1 — OPENING Use the opening composition and camera language of @Video 1 as the foundation. Establish @Image 1 naturally within the scene. Surrounding characters visible in @Video 1 may remain naturally present. @Image 1 begins her royal announcement, looking toward the LEFT side of the frame. Dialogue: "Shatru ko maarne wala yoddha hai." Her expression is calm, thoughtful and authoritative. Follow the camera behavior of @Video 1 while adapting it naturally to her performance. CUT / TRANSITION naturally according to the reference video's cinematic rhythm. SHOT 2 — CONTINUATION Create a NEW composition based on the next visual progression of @Video 1. Keep the surrounding characters and environment naturally integrated wherever visible. @Image 1 continues: "Praja ke liye apni vijay bhi tyaag de, wahi asli raja hai." Her expression gradually becomes more certain and emotionally powerful. Her gaze naturally shifts toward the RIGHT side of the frame. The camera responds naturally to her performance while maintaining @Video 1's visual language. Use realistic eye-lines and natural acting. SHOT 3 — FINAL ANNOUNCEMENT Continue the new cinematic progression inspired by @Video 1. Keep the environment and visible background characters naturally integrated. Bring the visual emphasis strongly back onto @Image 1. She looks firmly toward the RIGHT side and delivers: "Isiliye, Bhallaladev ko Mahishmati ka raja ghoshit kiya jata hai." The declaration must feel authoritative, decisive and emotionally powerful. After completing the sentence, hold her expression briefly. Her face communicates absolute certainty and royal authority. CAMERA LANGUAGE: @Video 1 remains the primary cinematic reference. Use its visual language and motion behavior as the foundation, intelligently adapted to the new dialogue. Use cinematic multi-shot progression, motivated camera movement, natural reframing, realistic transitions, appropriate close-ups, medium shots, wider compositions when required and natural depth of field. Avoid random camera movement, unnecessary zooms, spinning camera, artificial motion, exaggerated slow motion and generic AI cinematic movement. Every camera movement must support @Image 1's dialogue and emotional performance. DIALOGUE PRIORITY: The COMPLETE Hindi dialogue must be spoken EXACTLY as written. Do NOT translate, paraphrase, shorten or replace anything. Allow realistic pauses between sentences. Lip synchronization must be accurate. AUDIO — STRICTLY DIEGETIC: Only @Image 1's dialogue, realistic environmental ambience, natural character sounds, cloth movement, jewelry movement, footsteps, subtle crowd ambience and physically appropriate environmental SFX. ABSOLUTELY NO BGM. NO background music. NO orchestral score. NO emotional music. NO cinematic music. NO trailer music. NO musical buildup. VISUAL QUALITY: Premium photorealistic live-action Indian historical epic. Realistic human skin, natural facial texture, realistic eyes and hair, detailed historical costume and jewelry, natural fabric movement, physically accurate lighting, natural atmospheric depth and high-end theatrical cinematography. ENVIRONMENT: Use @RoyalCourt as the PRIMARY ENVIRONMENT REFERENCE. Integrate its royal court setting, architecture, lighting and atmosphere naturally into every shot while maintaining the cinematic language of @Video 1. FINAL PRIORITY: 1. @Image 1's realistic acting 2. Complete Hindi dialogue and accurate lip sync 3. Natural emotional progression 4. @Video 1's cinematic language and movement 5. Realistic environment and supporting characters 6. Premium photorealistic quality The result must feel like a NEW professionally directed historical-epic scene created using @Video 1 as the cinematic blueprint, with @Image 1's complete royal announcement as the central focus. No subtitles. No text. No logos. No watermark.

1 other text field— sigmas, filenames, and other non-prompt strings the logger swept up

node 2 · LoadImage · image

siva side.jpg

Input media

1

Source images and video this run consumed.

siva side.jpg

node 2 · 384 KB