Directorio de recursos
Encuentra lo que necesita tu flujo.
Guides: techniques, setup, and troubleshootingBusca Pipes de Sweet Tea, flujos externos, modelos, paquetes de nodos y guías en un catálogo con evidencias.
Empieza con un flujo ↗Resultados de búsqueda
Todos los recursos
18 mostrados of 41.761
Kling 3.0 first and last frame video
Generate one video between two reference images with Kling 3.0, with optional generated audio. Upload both a first frame and a last frame; both are required. Replace the example goldfish prompt with a description of the action, transition and camera movement connecting your images. For a first run, keep the five-second duration and 720p resolution, and leave output file type and codec set to auto. Duration supports 3–15 seconds. Update any timing written in the prompt when changing duration. Resolution offers 720p, 1080p and 4k. Generate audio is enabled by default: describe the desired ambience, music or dialogue in the prompt, or disable it when sound is unnecessary. Advanced Video output settings let you choose the file type and codec for the returned video. Generation sends the reference images and prompt to an authenticated partner API and incurs provider charges. Duration, resolution and audio settings affect the charge. The seed default of -1 requests fresh randomization each run. Seed changes control re-running, but a fixed seed does not guarantee identical results.
Grok image restyling
Upload one image and describe the appearance you want. For a first run, keep the pixel-art instruction, 1K resolution and one result. The prompt is fully editable: specify the desired style and which subjects or details should remain recognizable. Identity preservation is a request to the generative model, not a guarantee. The source contains equivalent style examples. This tool uses one editing branch with one editable prompt rather than running every example simultaneously. Copy any of these source starting instructions into the prompt: - Change the images style into pixel art, maintain identity - Change the images style into 3d animation, maintain identity - Change the images style into claymation, maintain identity - Change the images style into hyper realistic, maintain identity - Change the images style into cross stitch, maintain identity - Change the images style into oil painting with visible brushstrokes, maintain identity - Change the images style into detailed colored pencil, maintain identity - Change the images style into flat vector illustration, maintain identity - Change the images style into reflective metal sculpture, maintain identity - Change the images style into charcoal sketch, maintain identity - Change the images style into lineart, maintain identity - Change the images style into watercolor, maintain identity Choose 1K or 2K resolution and request 1–10 edited images. Model selection exposes the source-supported grok-imagine-image, grok-imagine-image-quality and grok-imagine-image-pro options. Aspect ratio remains automatic for this single-image input; the source permits explicit ratios only with multiple input images. Increasing result count does not create multiple input images. Seed controls whether the provider node re-runs; results remain nondeterministic even with a fixed seed. The default randomizes on each run. Outputs are generated raster still images: animation and sculpture instructions describe appearances, vector instructions do not produce vector files, and cross-stitch instructions do not produce stitch patterns. Pixel art is prompted without enforced grid or palette constraints. Uploaded transparency is not preserved by the selected image-only path.
Kling image generation
Generate images from text with Kling Image O1 or Kling 3.0 Omni and view the results in Sweet Tea. For a first run, keep Kling Image O1, 1K resolution, square aspect ratio, and image series disabled. Use the supplied fashion portrait prompt or replace it with your own subject, composition, lighting, and style. Describe unwanted elements in the same prompt when needed; there is no separate negative prompt. Choose the aspect ratio for image proportions and 1K, 2K, or 4K resolution for the requested output size. To request a series of 2–9 images, select Kling 3.0 Omni and then choose the image series count. Keep image series disabled with Kling Image O1, which does not support that mode. This Pipe accepts text only; no reference image input is connected. Leave seed at -1 to randomize the run seed. Seed controls whether generation re-runs, rather than reproducibility: results remain non-deterministic even with a fixed seed. Generation sends the prompt to the paid Kling provider. Results are delivered through Sweet Tea's image output; the workflow does not permanently save image files.
[LuisaP] Sci-fi/Hard surface
Needs the keyword " structure " ex: a cyberpunk structure , a building structure , a turret structure .
Hunyuan3D texture from images
Generate a texture image by reconstructing an object with Hunyuan3D through the provider API. For a first run, upload a front view of the object and leave Back view bypassed. Keep model 3.1, 500,000 faces and PBR enabled, then run the Pipe. To include a back view, enable Back view and upload a matching image of the same object. Model version offers Hunyuan3D 3.1 or 3.0. Face count requests between 3,000 and 1,500,000 mesh faces; it does not set texture pixel resolution. PBR requests physically based material generation, but this Pipe returns only the texture image. Seed -1 requests a fresh run. The provider produces non-deterministic results even when a seed is reused; the seed controls node re-running and caching. Generation is fixed to Normal for the texture output. Geometry mode does not supply this deliverable, and LowPoly requires a polygon-type control absent from this interface and is unavailable with model 3.1. This Pipe runs the underlying 3D generation on each requested generation run. It does not accept an existing mesh. It delivers no GLB or OBJ model, rendered model preview, or separate metallic, normal or roughness maps.
Wan 2.2 four-step frame-to-video
Generate a silent video from a scene and motion prompt, optionally guided by a start image, an end image, or both. For a first run, upload a start image and replace the sleeping-cat prompt with your subject, action, camera movement and style. Upload an end image if you want ending-frame guidance; otherwise disable the End image stage. Both image stages initially remain enabled, so upload an image for each enabled stage or disable it explicitly. Leave the default 640 × 640 dimensions, 81 frames, 16 FPS and sampling settings for the first run, then generate and view the returned video. The default playback lasts approximately 5.1 seconds. The image stages can be disabled independently: start only animates from an opening image, end only provides ending guidance, both guide a transition, and disabling both leaves prompt-driven generation. Images provide guidance rather than a guarantee of an exact transition. Only RGB image outputs are connected; uploaded alpha masks are unused. There is no audio input in this tool. Width and height are pixels in increments of 16. Frame count follows 1, 5, 9 and so on and changes how many frames are generated. Playback FPS changes duration without generating additional frames: duration is frame count divided by FPS. Keep batch size at 1 for a single video; this Pipe has one video output rather than separate outputs per batch item. Advanced settings preserve separate initial and finishing-stage controls. Seed -1 randomizes initial noise; a fixed nonnegative seed repeats the noise choice with the same other settings. The high-noise stage adds noise and runs steps 0–2, then the low-noise stage continues steps 2–4 without adding fresh noise. Both use a total schedule length of 4, CFG 1, Euler, simple scheduling and sampling shift 8. The fixed high/low-noise Lightx2v LoRAs each have model strength 1. If deliberately changing this recipe, coordinate both total step counts and the segment boundaries; changing one field alone can interrupt the handoff. LoRA strengths do not select the standard 20-step sibling recipe. The supplied negative prompt is a Chinese list of unwanted visual defects and content; retain it or edit it for your scene. Output container, codec, CRF and bit depth control video representation and encoding. Leave container and codec on auto and CRF at -1 for the first run. Ten-bit output may be unsupported by some players or downstream nodes.
Wan 2.6 text to video
Generate a video from an English or Chinese scene description using Wan, with optional generated audio and continuous-shot or multi-shot composition. No uploaded image, video, or audio input is exposed in this Pipe. For a first run, replace the example scene description with your subjects, actions, setting, camera movement, and visual style. Keep Wan 2.6, landscape 720p, and the five-second duration. Audio generation and AI prompt enhancement are enabled by default; include desired sound in your description when using generated audio. Review the negative prompt and replace scene-specific exclusions that do not apply. Output size selects resolution and aspect ratio together. Duration is in seconds: Wan 2.6 supports 5, 10, or 15; when selecting Wan 2.5 preview, choose 5 or 10. Shot type selects one continuous shot or multiple shots with cuts, and takes effect only while Enhance prompt is enabled. Disable enhancement to submit your wording without AI expansion. Turn off Generate audio if you do not want automatic audio generation. Seed -1 randomizes each run; enter a fixed seed to control randomization. Watermark requests an AI-generated watermark and is disabled by default. The generated video is returned to Sweet Tea. In the advanced Video output section, leave file type and codec on auto for the first run, or select an explicit container and encoding. Keep the supplied CRF default unless you need to configure encoding quality.
Ideogram V4 image generator
Generate an image through the Ideogram V4 partner API using a written prompt. No source image is required. For a first run, keep the bundled surreal collage-poster prompt, Auto output size, and DEFAULT rendering speed, then run. The execution environment needs access to the partner API. Replace the prompt with plain natural language for a simple idea, or edit the bundled JSON to describe composition, colors, objects, and in-image wording. JSON remains one prompt field; this Pipe does not provide separate per-object editors or a negative prompt. Treat layout and lettering instructions as generation requests rather than guaranteed exact placement or transcription. Choose an explicit output size when you need a particular pixel size and aspect ratio, and keep composition instructions consistent with that shape. Rendering speed offers DEFAULT, TURBO, and QUALITY modes for the provider's speed-versus-quality trade-off. Seed defaults to a new random value each run. The source's Gemini helper is disabled and disconnected. Prompt building is manual in this Pipe; it does not automatically expand ideas into JSON.
Action figure image edit
Create an articulated toy or action-figure style image from an uploaded image. Describe the edit in the positive prompt using "action the [thing]", such as "action the scene"; optionally add exclusions in the negative prompt. Fast mode selects the Lightning LoRA branch with four sampling steps by default; disabling it selects standard sampling. Advanced controls adjust edit strength, LoRA strengths, seed, sampler, scheduler, and each mode's steps and guidance.
Product placement video
Create a product-placement video from three product photos and one composition reference. Nano Banana generates a new start frame from all four images; Gemini examines it and writes a motion prompt; Kling animates it. Every run includes composition generation. This tool does not accept an existing finished start frame or return the intermediate still as a separate output. For a first run, upload all four required images: products are images 1–3, and the composition reference is image 4. Edit Placement prompt to describe how the products should appear in that scene. Its default assumes products held in a hand in a fanned arrangement; revise it for other compositions. In Motion instructions, describe desired subject movement and camera motion. These instructions go to Gemini, whose response becomes Kling's actual prompt automatically. Keep the initial 1K start frame, 5-second duration, 1080p video, and automatic output encoding settings. Start frame aspect ratio determines the composition shape and the resulting video's framing. Auto follows an input image's aspect ratio, but the supplied source does not identify which of the four takes precedence. Choose an explicit ratio when framing matters. Kling's own aspect-ratio field is ignored with the generated start frame connected. Start frame resolution and Video resolution are independent. The three model selectors act independently. Image model offers Nano Banana 2 (Gemini 3.1 Flash Image) and Nano Banana 2 Lite. Both retain all four image bindings and share aspect-ratio and thinking-level controls. For Start frame resolution, enter 1K, 2K, or 4K with the full model; set 1K before switching to Lite, which supports only 1K. Thinking level offers MINIMAL or HIGH. Image temperature controls randomness from 0 to 2, and Image nucleus sampling controls the sampling threshold from 0 to 1; lower values favor more focused choices. Keep their supplied defaults initially. Image generation instructions changes foundational behavior; use Placement prompt for ordinary scene changes and preserve image production. Motion writer model offers gemini-2.5-pro, gemini-2.5-flash, gemini-3-pro-preview, gemini-3-1-pro, and gemini-3-1-flash-lite. Motion writer instructions optionally changes its foundational behavior. Leave that field empty initially; any resulting response is consumed directly as the video prompt. Video model offers kling-v3 and kling-3.0-turbo. kling-v3 supports 4k, 1080p, and 720p; Turbo supports 1080p and 720p, so switch away from 4k before selecting Turbo. Generate audio controls kling-v3. Turbo always generates native audio and ignores this toggle. The retained mode generates a single shot lasting 3–15 seconds; storyboard controls are not exposed. Negative prompt describes unwanted elements for video generation and does not affect the still composition. A seed of -1 randomizes each run. Fixed image and motion-writer seeds request best-effort repeatability without guaranteeing identical results. Kling's seed controls re-running rather than reproducible video generation. File type and Codec affect the returned video's encoding; keep auto initially. Inspect product identity, packaging, label text, pose, and motion before using the result: prompt requests are not guarantees of exact reproduction. The upload loaders' alpha-mask outputs are not connected, so uploaded transparency is not preserved through a separate mask path. Text and thinking-image outputs from image generation are not returned. For a still composition without motion writing or animation, use the separate Product placement photo tool.
Hentai Manga Style Concept
This is a embedding i trained on just hentai images. Therefore you will get a lot of open mouths, ahegao, sweating, saliva etc.. from just using the embedding.
Topaz video upscale
Enhance an uploaded video with Topaz Starlight Precise 2.5 and return one video at FullHD (1080p) or 4K (2160p). For a first run, upload your source video, keep FullHD (1080p), leave Compression level at Low, and leave File type and Codec on auto. Run the Pipe to receive the enhanced video. A source video is required; no sample video is supplied as the upload default. Output resolution selects the Topaz enhancement resolution. Compression level selects Topaz's CQP setting from Low, Mid and High. In Video encoding, you can choose MP4, MKV or WebM and H.264 or AV1 instead of automatic selection. Manual file type and codec combinations must be compatible and supported by the runtime. These output encoding choices are separate from Topaz's compression setting. This Pipe uses Starlight Precise 2.5 with frame interpolation disabled. Other enhancement models, creative-model tuning, target frame rate and slow-motion settings are not offered because their required mode-specific bindings are absent from this configuration. Processing sends the uploaded video to the Topaz partner API, requires runtime authentication, and uses provider-priced processing.
Qwen illustration
Generate illustrations from text using Qwen-Image with the fixed Alvdansen Illustration 1.0 LoRA. Describe the subject, composition, colors and desired style in Image prompt, or keep the included example for a first run. Optionally list unwanted elements in Negative prompt; it can remain empty. No uploaded image is required or supported. For a first run, keep the 1024 × 1024 dimensions, image count 1 and illustration strength 1. The advanced sampling defaults are 45 steps, prompt guidance (CFG) 3.5, Euler sampler, Simple scheduler, denoising amount 1 and sampling shift 7. Width and height are pixels and must be multiples of 16. Image count sets the batch size; larger dimensions and batches increase resource use. The output returns every decoded image in the batch through Sweet Tea's image preview transport. Illustration style is enabled by default. Reduce its strength for less LoRA influence, or disable that stage to generate with the base Qwen-Image model. Illustration strength has no effect while the stage is disabled; prompts, output dimensions and sampling controls still apply. Seed -1 selects fresh starting noise each run. Use a fixed seed when comparing changes with the same starting noise. Advanced controls adjust denoising steps, prompt guidance, the sampling algorithm and noise schedule. Keep denoising amount at 1 initially: lowering it does not preserve an input image because generation starts from an empty latent. The model, text encoder, VAE and illustration LoRA are fixed in this tool.
myself-badhand
This embedding focuses on correcting issues related to the number and shape of fingers, which may result in a change in the model's style.
HTRP Mesugaki/Gyaru
This isn't a character model; it's a concept. When combined with LORA, the character becomes a fusion of gyaru and mesugaki.
Z-Image Turbo edge-guided generation
Generate images from a prompt while following a reference image's Canny edges. Each run returns generated images and a preview of the reference edge map. For a first run, upload a reference with the composition and outlines you want, then enter a prompt describing the desired subject, appearance, lighting and style. Keep control strength at 1, image count at 1, and the source sampling setup: 8 steps, CFG 1, res_multistep sampler, simple scheduler and denoise amount 1. Seed -1 provides fresh randomization; use a fixed seed to compare changes with the same initial noise. Inspect the returned edge preview to understand the structural guidance. The normalized low and high Canny thresholds default to 0.1 and 0.32. Lower thresholds retain more faint detail; higher thresholds suppress weaker edges. Keep the low threshold below the high threshold. Control strength scales the edge guidance influence; reducing it toward 0 loosens that influence. Output dimensions are derived from the processed reference, with no independent width or height controls. Resize reference is bypassed by default. Enable it to choose a longest side in pixels, initially 1024, while retaining the reference aspect ratio. The resize method affects the image used for edge extraction. Resizing changes both the guidance map and the dimensions supplied to generation. Increase image count for a batch of variations; the edge preview remains one reference map. Advanced sampling controls include steps, CFG, sampler, scheduler, denoise amount and sampling shift. Their source defaults provide the initial setup, and changes affect sampling behavior. The reference is structural guidance, not an encoded starting image: its original colors, textures and transparency are not preserved. Lower denoise does not turn this into pixel-preserving image editing. Negative conditioning is zeroed internally, so there is no editable negative prompt. This Pipe implements Canny guidance only.
Corneo's Power (Chainsaw Man) TI Embedding
This is a textual inversion embedding for Power from Chainsaw Man. Chainsaw Man is an incredible series that you should try out if you haven't already.
Nano Banana 2 image generation
Generate images from a text description through the Nano Banana 2 partner API. Configured partner API access is required; requests send your prompt and settings to the provider and use paid provider generation. This Pipe accepts text only, with no connected reference-image input, and returns the final image output for preview. For a first run, use the supplied futuristic poster prompt or describe your own subject, composition, style, and exact lettering. Start with the default Nano Banana 2 model, square Aspect ratio, 1K Resolution, and MINIMAL Thinking level. Include unwanted elements as instructions in Prompt; there is no separate negative prompt. Keep proportions mentioned in the prompt consistent with Aspect ratio. Choose an explicit ratio for a predictable requested shape; auto leaves the shape to the provider. Choose the model before setting Resolution. Enter 1K, 2K, or 4K for Nano Banana 2, or 1K for Nano Banana 2 Lite. Resolution is a text field, so enter these values exactly. Actual dimensions also depend on Aspect ratio. Thinking level offers MINIMAL and HIGH reasoning; thinking images are not returned. Seed -1 randomizes each run. A fixed seed requests best-effort consistency, but identical images are not guaranteed. Temperature ranges from 0 to 2 and controls randomness; Top p ranges from 0 to 1 and controls the sampling threshold. Lower values favor more focused responses, while higher values allow more diversity. System instructions set the provider's foundational behavior; the supplied default directs requests toward image generation. Response format can request IMAGE or IMAGE+TEXT. Only the final image output is connected to this Pipe's result, so accompanying text is not returned. The output stage provides a preview rather than the source workflow's SaveImage disk-saving step.
