The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.

文件	最后提交记录	最后更新时间
.ci	Use windows line endings for windows portable readmes. (#14334)	2 个月前
.github	Remove some workflow media from portable install to make it smaller. (#16080)	4 天前
alembic_db	feat(assets): add namespaced model_type tags and align tag semantics (#14511) * feat(assets): add namespaced model type tags * fix(assets): mark path-derived upload tags automatic * fix(assets): merge duplicate scan specs * test(assets): make duplicate path normalization portable * feat(assets): add loader_path as the authoritative loader locator (#14796) * fix(assets): filter model_type tags by bucket extension sets Buckets sharing a base directory (e.g. diffusion_models and a custom unet_gguf) tagged every file in the directory regardless of whether the bucket could load it, so .safetensors files were tagged model_type:unet_gguf and vice versa. Carry each bucket's registered extension set through get_comfy_models_folders and only emit a model_type tag when the file extension matches, keeping the empty-set match-all convention from folder_paths.filter_files_extensions. Files under a model base matching no bucket now keep only the models tag instead of every directory-matching model_type tag. * feat(assets): replace response file_path with persisted loader_path The old file_path response field was a namespaced storage locator (models/checkpoints/foo.safetensors): not an absolute path, not unique identity, and not the value a loader consumes. Nothing needs that shape on the wire (hash/ID-based locating is the long-term direction), so it is dropped rather than renamed; the storage-root matching stays internal, powering display_name. What loaders DO need is the in-root loader path (category dropped: models/checkpoints/foo/bar.safetensors -> foo/bar.safetensors). Serve it as a first-class loader_path field, persisted on asset_references (migration 0006) and written by every ingest pipeline at insert, so responses read the column verbatim. Like the model_type tags, loader_path is a seed-time derivative of the model folder registry, maintained by the same scan lifecycle (new files seed fresh values, pruning retires rows whose bucket disappeared). Rows predating the column serve a null loader_path; databases from before this stack already need recreating for the base branch's tag changes. loader_path resolves every registered base including extra_model_paths entries; display_name only the canonical storage roots. A file can therefore be loadable with no display name (extra-path models) or the reverse (unregistered files under the models root), and loader_path is null exactly when no loader can resolve the file. * test(assets): lock loader_path matrix (asymmetry, null, persist/read) Cover the behaviour that has no production change but is easy to regress: the extra-path asymmetry (loadable but no storage namespace), null loader_path persistence for orphan files, and the response reading the stored column with a compute fallback for un-backfilled rows. * fix(assets): persist subfolder-qualified loader_path for ingested outputs ingest_existing_file built its seed spec with the file's basename, so outputs saved into a subfolder persisted loader_path (and the user_metadata filename that preview URLs split for their subfolder param) as just the basename: the served locator pointed at a file that does not exist at that path. Scanner and seeder specs already derive fname via compute_loader_path; use the same derivation here. * fix(assets): only extension-matching buckets contribute a loader_path The model-base match in get_asset_category_and_relative_path ignored each bucket's extension set, so a file inside a registered base whose extension the bucket cannot load (e.g. a .txt uploaded into model_type:checkpoints) advertised a loader_path that no loader list would ever resolve, while the tag side of the same stack already excluded it. Apply the extension check used for backend tags (empty set accepts any extension), keeping loader_path null exactly when no loader can resolve the file. * fix(assets): refresh loader_path when re-ingesting an existing reference upsert_reference only wrote loader_path on the INSERT branch, so re-ingesting an existing reference (an output overwritten in place, or a file re-registered after its loader_path derivation changed) kept the stale or NULL value forever. Write it on the UPDATE branch too, with a null-safe change guard so a loader_path difference alone is enough to trigger the update, and identical values stay a no-op. * fix(assets): repair semantic merge breakage from #14796 and master Two textually-clean but semantically-broken merges: - routes.py lost its folder_paths import when #14796's import block superseded the base's, while the content-type hardening added via the base's master merge still calls folder_paths.is_dangerous_content_type. - master's SVG download-hardening test uploads with the pre-namespacing bare checkpoints tag, which this branch's destination validation rejects; use model_type:checkpoints. --------- Co-authored-by: guill <jacob.e.segal@gmail.com>	1 个月前
api_server	fix: append directory type annotation to internal files endpoint response (#13078) (#13305)	4 个月前
app	Make comfyui handle missing workflow media properly. (#16079)	4 天前
blueprints	Drop the duplicate first image from the Qwen-Image-Layered blueprint (#15364)	20 天前
comfy	Pause comfy compiler for long lived sparse allocations (#16148) * model_prefetch: add friendly API to pause comfy compiler * sparse_attention: Pause comfy compiler for un-freed pieces The VSA plan and pool stats do not free within the transformer causing a comfy-compiler rogue allocation and graph break. Ideally these should be freed in the lifetime of the transformer run (and comfy-compiler scope) but there isnt a convenient place for the particular lifecylce intended by these variables. So go with the pause() approach.	1 天前
comfy_api	Lower peak trellis workflow vram and ram usage. (#16054)	4 天前
comfy_api_nodes	[Partner Nodes] deprecate retired models (#16121) * [Partner Nodes] chore(OpenAI): remove the DALL·E 2 and DALL·E 3 nodes, OpenAI shut both models down Signed-off-by: Alexander Piskun <bigcat88@icloud.com> * [Partner Nodes] chore(LTX): remove the LTX-2 nodes, the vendor no longer serves ltx-2-fast and ltx-2-pro Signed-off-by: Alexander Piskun <bigcat88@icloud.com> * [Partner Nodes] chore(ByteDance): remove the Seedream 3.0 node and the Seedance 1.0 Lite models, BytePlus deactivated them Signed-off-by: Alexander Piskun <bigcat88@icloud.com> * [Partner Nodes] chore(Kling): remove the Video Extend node, video-extend only accepted videos from the retired 1.x models Signed-off-by: Alexander Piskun <bigcat88@icloud.com> * [Partner Nodes] chore(ByteDance): remove the Reference Images to Video node, no Seedance 1.0 model accepts reference images Signed-off-by: Alexander Piskun <bigcat88@icloud.com> --------- Signed-off-by: Alexander Piskun <bigcat88@icloud.com>	2 天前
comfy_config	Add new fields to the config types (#8507)	1 年前
comfy_execution	Lower peak trellis workflow vram and ram usage. (#16054)	4 天前
comfy_extras	Don't blend the alpha channel in the Blend Images node (#15625)	1 天前
custom_nodes	Update nodes categories and display names (CORE-89) (#13786)	3 个月前
input	LoadLatent and SaveLatent should behave like the LoadImage and SaveImage.	3 年前
middleware	fix: use no-store cache headers to prevent stale frontend chunks (#12911) After a frontend update (e.g. nightly build), browsers could load outdated cached index.html and JS/CSS chunks, causing dynamically imported modules to fail with MIME type errors and vite:preloadError. Hard refresh (Ctrl+Shift+R) was insufficient to fix the issue because Cache-Control: no-cache still allows the browser to cache and revalidate via ETags. aiohttp's FileResponse auto-generates ETags based on file mtime+size, which may not change after pip reinstall, so the browser gets 304 Not Modified and serves stale content. Clearing ALL site data in DevTools did fix it, confirming the HTTP cache was the root cause. The fix changes: - index.html: no-cache -> no-store, must-revalidate - JS/CSS/JSON entry points: no-cache -> no-store no-store instructs browsers to never cache these responses, ensuring every page load fetches the current index.html with correct chunk references. This is a small tradeoff (~5KB re-download per page load) for guaranteed correctness after updates.	5 个月前
models	Update MediaPipe nodes to standardize with existing code base (CORE-242) (#14025)	3 个月前
output	Initial commit.	3 年前
script_examples	Update comment in api example. (#9708)	11 个月前
tests-unit	Don't blend the alpha channel in the Blend Images node (#15625)	1 天前
tests	Keep temp-directory assets visible while their files exist (#15510) * Keep temp-directory assets visible while their files exist Assets written to the temp directory were flagged as missing and dropped from GET /api/assets, even with the file sitting on disk. One list of directories was answering two different questions -- where the scanner looks for new files, and which files ComfyUI considers its own -- and temp belongs only in the second, so every temp reference was disowned by the prune that runs at startup and on POST /api/assets/prune. Ownership now covers temp. Discovery still does not: the temp directory is wiped before the scan runs, and assets written there are already registered with a hash, mime type and dimensions, so walking it would find nothing. Temp references are instead reconciled against the filesystem directly, so a temp file that really is gone is still retired rather than lingering as a broken entry. get_prefixes_for_root becomes get_scan_prefixes_for_root so the two questions are told apart by name rather than by comment. * Cover the unhashed temp asset in the reconciliation tests The existing temp tests all registered hashed assets, so they never exercised the path an unhashed asset takes when its file is gone: the orphaned rows are removed rather than kept as missing, exactly as under any other root.	21 天前
utils	Update logging level for invalid version format (#13526)	4 个月前
.coderabbit.yaml	chore: fail the CodeRabbit status when the review did not run (#15382)	1 个月前
.gitattributes	Add Veo3 video generation node with audio support (#9110) - Create new Veo3VideoGenerationNode that extends VeoVideoGenerationNode - Add support for generateAudio parameter (only for Veo3 models) - Support new Veo3 models: veo-3.0-generate-001, veo-3.0-fast-generate-001 - Fix Veo3 duration constraint to 8 seconds only - Update original node to be clearly Veo 2 only - Update API paths to use model parameter: /proxy/veo/{model}/generate - Regenerate API types from staging to include generateAudio parameter - Fix TripoModelVersion enum reference after regeneration - Mark generated API types file in .gitattributes	1 年前
.gitignore	Add deploy environment header (Comfy-Env) to partner node API calls (#13425)	4 个月前
.spectral.yaml	Suppress false-positive Spectral lint on WebSocket endpoint (#13842) The /ws path uses HTTP 101 (Switching Protocols), which is the correct response for a WebSocket upgrade but not a 2xx. The built-in operation-success-response rule fires as a false positive because OpenAPI 3.x has no native WebSocket support. Add a path-scoped override in .spectral.yaml to disable the rule for /ws only, leaving it active for all other operations.	3 个月前
AGENTS.md	Add how to cache model stuff to AGENTS.md (#16003)	7 天前
CODEOWNERS	chore: add /AGENTS.md to CODEOWNERS (#14962) Scope AGENTS.md review to @comfyanonymous, matching the existing /CODEOWNERS, /.ci/, and /.github/ meta-file entries.	1 个月前
CONTRIBUTING.md	Add CONTRIBUTING.md (#3910) * Create CONTRIBUTING.md * Add feature-request channel link. * Remove discord links for channels.	2 年前
LICENSE	Initial commit.	3 年前
QUANTIZATION.md	Update quant doc so it's not completely wrong. (#13381) There is still more that needs to be fixed.	4 个月前
README.md	Update README.md (#16071) ComfyUI Desktop has proven to be the easiest way to install and manage local ComfyUI instances. We will still maintain a portable version for now but this change is to nudge the community in using Desktop instead.	4 天前
SECURITY.md	Create SECURITY.md. (#13902)	3 个月前
alembic.ini	Add support for sqlite database (#8444) * Add support for sqlite database * fix	1 年前
comfyui_version.py	ComfyUI v0.34.0	13 天前
cuda_malloc.py	Limit Windows multi-GPU visibility (#15737)	19 天前
execution.py	Introduce Comfy Compiler (CORE-389) (#15861)	3 天前
extra_model_paths.yaml.example	[Dataset/Security,Feature] Add dataset folder to avoid arbitrary folder access for dataset stuff. (#14807)	1 个月前
folder_paths.py	Use the actual function to check if the path is within the dir. (#15216)	1 个月前
hook_breaker_ac10a0.py	Prevent custom nodes from hooking certain functions. (#7825)	1 年前
latent_preview.py	Add support for taeh3. (#15695)	21 天前
main.py	Add warning about forcing single GPU mode on Windows (#16055)	5 天前
manager_requirements.txt	bump manager version to 4.2.2 (#14471)	2 个月前
node_helpers.py	Fix issue blend images with alpha (#13615) Make ImageBlend and ImageCompositeMasked nodes handle images with different channel counts	4 个月前
nodes.py	Add Sparse Attention node (#16072) * support VSA * BlockSparseAttention node * Add extra_tokens * cleanup * cleanup	2 天前
openapi.yaml	chore(openapi): sync shared API contract from cloud@94d0f1b (#15041) Co-authored-by: mattmillerai <7741082+mattmillerai@users.noreply.github.com> Co-authored-by: Matt Miller <mattmiller@comfy.org>	25 天前
protocol.py	Support for async node functions (#8830) * Support for async execution functions This commit adds support for node execution functions defined as async. When a node's execution function is defined as async, we can continue executing other nodes while it is processing. Standard uses of `await` should "just work", but people will still have to be careful if they spawn actual threads. Because torch doesn't really have async/await versions of functions, this won't particularly help with most locally-executing nodes, but it does work for e.g. web requests to other machines. In addition to the execute function, the `VALIDATE_INPUTS` and `check_lazy_status` functions can also be defined as async, though we'll only resolve one node at a time right now for those. * Add the execution model tests to CI * Add a missing file It looks like this got caught by .gitignore? There's probably a better place to put it, but I'm not sure what that is. * Add the websocket library for automated tests * Add additional tests for async error cases Also fixes one bug that was found when an async function throws an error after being scheduled on a task. * Add a feature flags message to reduce bandwidth We now only send 1 preview message of the latest type the client can support. We'll add a console warning when the client fails to send a feature flags message at some point in the future. * Add async tests to CI * Don't actually add new tests in this PR Will do it in a separate PR * Resolve unit test in GPU-less runner * Just remove the tests that GHA can't handle * Change line endings to UNIX-style * Avoid loading model_management.py so early Because model_management.py has a top-level `logging.info`, we have to be careful not to import that file before we call `setup_logging`. If we do, we end up having the default logging handler registered in addition to our custom one.	1 年前
pyproject.toml	ComfyUI v0.34.0	13 天前
pytest.ini	Execution Model Inversion (#2666) * Execution Model Inversion This PR inverts the execution model -- from recursively calling nodes to using a topological sort of the nodes. This change allows for modification of the node graph during execution. This allows for two major advantages: 1. The implementation of lazy evaluation in nodes. For example, if a "Mix Images" node has a mix factor of exactly 0.0, the second image input doesn't even need to be evaluated (and visa-versa if the mix factor is 1.0). 2. Dynamic expansion of nodes. This allows for the creation of dynamic "node groups". Specifically, custom nodes can return subgraphs that replace the original node in the graph. This is an incredibly powerful concept. Using this functionality, it was easy to implement: a. Components (a.k.a. node groups) b. Flow control (i.e. while loops) via tail recursion c. All-in-one nodes that replicate the WebUI functionality d. and more All of those were able to be implemented entirely via custom nodes, so those features are not a part of this PR. (There are some front-end changes that should occur before that functionality is made widely available, particularly around variant sockets.) The custom nodes associated with this PR can be found at: https://github.com/BadCafeCode/execution-inversion-demo-comfyui Note that some of them require that variant socket types ("") be enabled. Allow `input_info` to be of type `None` * Handle errors (like OOM) more gracefully * Add a command-line argument to enable variants This allows the use of nodes that have sockets of type '' without applying a patch to the code. Fix an overly aggressive assertion. This could happen when attempting to evaluate `IS_CHANGED` for a node during the creation of the cache (in order to create the cache key). * Fix Pyright warnings * Add execution model unit tests * Fix issue with unused literals Behavior should now match the master branch with regard to undeclared inputs. Undeclared inputs that are socket connections will be used while undeclared inputs that are literals will be ignored. * Make custom VALIDATE_INPUTS skip normal validation Additionally, if `VALIDATE_INPUTS` takes an argument named `input_types`, that variable will be a dictionary of the socket type of all incoming connections. If that argument exists, normal socket type validation will not occur. This removes the last hurdle for enabling variant types entirely from custom nodes, so I've removed that command-line option. I've added appropriate unit tests for these changes. * Fix example in unit test This wouldn't have caused any issues in the unit test, but it would have bugged the UI if someone copy+pasted it into their own node pack. * Use fstrings instead of '%' formatting syntax * Use custom exception types. * Display an error for dependency cycles Previously, dependency cycles that were created during node expansion would cause the application to quit (due to an uncaught exception). Now, we'll throw a proper error to the UI. We also make an attempt to 'blame' the most relevant node in the UI. * Add docs on when ExecutionBlocker should be used * Remove unused functionality * Rename ExecutionResult.SLEEPING to PENDING * Remove superfluous function parameter * Pass None for uneval inputs instead of default This applies to `VALIDATE_INPUTS`, `check_lazy_status`, and lazy values in evaluation functions. * Add a test for mixed node expansion This test ensures that a node that returns a combination of expanded subgraphs and literal values functions correctly. * Raise exception for bad get_node calls. * Minor refactor of IsChangedCache.get * Refactor `map_node_over_list` function * Fix ui output for duplicated nodes * Add documentation on `check_lazy_status` * Add file for execution model unit tests * Clean up Javascript code as per review * Improve documentation Converted some comments to docstrings as per review * Add a new unit test for mixed lazy results This test validates that when an output list is fed to a lazy node, the node will properly evaluate previous nodes that are needed by any inputs to the lazy node. No code in the execution model has been changed. The test already passes. * Allow kwargs in VALIDATE_INPUTS functions When kwargs are used, validation is skipped for all inputs as if they had been mentioned explicitly. * List cached nodes in `execution_cached` message This was previously just bugged in this PR.	2 年前
requirements.txt	chore: update embedded docs to v0.5.11 (#15947) Co-authored-by: Alexis Rolland <alexisrolland@hotmail.com>	2 天前
server.py	Fix SVG previews broken by the stored-XSS forced-download (#15149) * Fix SVG previews broken by the stored-XSS forced-download /view and the assets download route force every SVG to application/octet-stream + attachment. That blocks the stored XSS from GHSA-779p-m5rp-r4h4, but it also breaks the SVG node output and Media Assets previews, which request the file with a plain <img>. Exempt only that case. An SVG referenced by an <img> loads in secure static mode with scripting and external references disabled, so the payload cannot fire. The attack needs the SVG to become a document, which arrives with a different Sec-Fetch-Dest. Browsers set that header themselves and page script cannot override it. A missing header, from a non-browser client or a proxy that strips it, fails closed. The blocklist itself is unchanged; this is a call-site gate. * Don't let a cache replay the inline SVG into document context The Sec-Fetch-Dest exemption makes /view and the assets content route vary their Content-Type and Content-Disposition on a request header, but neither response said so. FileResponse emits Last-Modified/ETag and the cache_control middleware skips /view (the filename is in the query string, not the path), so the inline image/svg+xml variant is heuristically cacheable. A cache keyed on the URL alone could hand an entry primed by an <img> load to a later top-level navigation of the same URL, turning the SVG back into a document and re-enabling the stored XSS the forced download blocks. Set Vary: Sec-Fetch-Dest and Cache-Control: no-store on both branches, not just the exempt one: a cached attachment replayed to an <img> would re-break the preview this fix exists to restore. Also strip parameters from content_type before building the assets response. mime_type there is uploader-supplied and unvalidated, and aiohttp rejects a charset in the content_type argument with ValueError, so a stored "image/svg+xml; charset=utf-8" turned a valid inline SVG into a 500. Route-level guards now pin the headers on both branches and the parameterised mime type; all three fail against the previous commit.	1 个月前

自动翻译

ComfyUI

最强大、高度模块化的内容创作 AI 引擎。

ComfyUI 是面向视觉专业人士的 AI 创作引擎，满足其对每个模型、每个参数和每个输出结果的掌控需求。其强大而模块化的节点图界面，帮助创作者生成图像、视频、3D 模型、音频等更多内容……

ComfyUI 原生支持最新的开源前沿模型。
合作伙伴节点可访问 Nano Banana、Seedance、Hunyuan3D 等顶级闭源模型。
它支持 Windows、Linux 和 macOS，可通过我们的桌面应用、便携安装在本地使用，或在我们的云服务上运行。
得益于 App Mode，最复杂的工作流也能通过简洁的 UI 直观呈现。
它可借助我们的 API 端点无缝集成到生产流水线中。

快速上手

本地

手动安装

支持所有操作系统与 GPU 类型（NVIDIA、AMD、Intel、Apple Silicon、Ascend）。

云端

示例

通过新版模板工作流或旧版示例工作流探索 ComfyUI 的能力。

功能特性

可视化节点图，无需代码即可构建并复用图像、视频、音频、3D 与文本工作流。
可复用子图、工作流模板、App Mode，以及用于将工作流集成到应用程序中的本地 API。
高效的本地执行，支持异步排队、部分图重新执行、智能 VRAM 与 RAM 管理、模型卸载以及量化模型。
广泛的原生模型支持。以下为代表性列表；可浏览工作流库查看持续维护、可直接运行的模板。
- 图像生成: Stable Diffusion 1.5、SDXL、SD3.5、Flux.1、Flux.2、Qwen Image、Z-Image、Hunyuan Image 2.1、HiDream、Lumina Image 2.0、Chroma、Anima、LongCat Image、Ideogram 4、Krea 2、MageFlow、Microsoft Lens、PixelDiT、Kandinsky 5 与 Ernie Image。
- 图像编辑: Flux Kontext、Flux.2 Klein、Qwen Image Edit、HiDream E1.1 与 O1、OmniGen2、Boogu、JoyImage Edit、MageFlow Edit 与 LongCat Image Edit。
- 视频生成: Wan 2.1 与 2.2、LTX-Video 2 与 2.3、HunyuanVideo 1.5、Kandinsky 5 Video、CogVideoX、Cosmos Predict2、Bernini-R、SCAIL 2 与 Mochi。
- 音频与视频生成: MiniMax H3 与 LTX-AV。
- 音频生成: ACE-Step 1.5、Stable Audio 3 与 MiniMax Music 3。
- 3D 与视觉: Hunyuan3D 2.1、TripoSplat、SeedVR2、SUPIR、Depth Anything 3、MoGe、SAM 3 与 3.1、RT-DETRv4 与 BiRefNet。
- 文本生成: Gemma 3 与 4、Qwen3、Qwen3.5 与 Qwen3-VL，包含多模态输入。
可从受支持的模型格式中加载完整检查点，或分别加载扩散模型、VAE、文本编码器、LoRA、ControlNet、适配器与超分模型。
内置局部重绘、扩展重绘、参考条件、蒙版与合成、模型合并、超分、帧插值、分割、深度估计与媒体处理工具。
支持以 JSON 格式保存与加载工作流，或从受支持的生成媒体中恢复完整工作流与随机种子。
完全离线运行：核心组件不会下载任何内容，除非你主动请求。使用 --disable-api-nodes 可禁用可选的付费 Comfy API 节点，并强制所有内置功能保持离线。
使用自定义节点扩展 ComfyUI。
通过 extra_model_paths.yaml 配置额外模型路径。
支持保存与加载高色深图像和视频：支持 16 位 PNG 图像、32 位 EXR、10 位 AVIF 等更多格式。
支持以多种格式保存与加载 HDR 视频和图像。

发布流程

ComfyUI 通常以每周发布一次的节奏推进，目标为周一，但因新模型发布或代码库重大调整而经常变动。共有三个相互关联的仓库：

ComfyUI Core
- 大约每 2 周发布一个新的主要稳定版本（例如 v0.7.0）。
- 从 v0.4.0 开始，补丁版本将用于将修复回移至当前稳定版本。
- 次要版本将用于基于 master 分支的发布。
- 在回移不合理的场景下，master 分支上的发布仍可能使用补丁版本。
- 稳定版本标签之外的提交可能非常不稳定，并可能使许多自定义节点失效。
- 作为桌面端发布的基础。
Comfy Desktop
- 使用最新稳定核心版本构建新的发布版本。
ComfyUI Frontend
- 每隔至少 2 周，前端更新会合并到核心仓库中。
- 为即将到来的核心版本发布冻结功能。
- 同时继续推进下一个发布周期的开发。

快捷键

快捷键	说明
`Ctrl` + `Enter`	将当前节点图加入生成队列
`Ctrl` + `Shift` + `Enter`	将当前节点图作为首个任务加入生成队列
`Ctrl` + `Alt` + `Enter`	取消当前生成
`Ctrl` + `Z`/`Ctrl` + `Y`	撤销/重做
`Ctrl` + `S`	保存工作流
`Ctrl` + `O`	加载工作流
`Ctrl` + `A`	选择所有节点
`Alt` + `C`	折叠/展开所选节点
`Ctrl` + `M`	静音/取消静音所选节点
`Ctrl` + `B`	旁路所选节点（效果如同节点从节点图中移除，连接线直接贯通）
`Delete`/`Backspace`	删除所选节点
`Ctrl` + `Backspace`	删除当前节点图
`Space`	按住并移动光标以平移画布
`Ctrl`/`Shift` + `Click`	将点击的节点加入选区
`Ctrl` + `C`/`Ctrl` + `V`	复制并粘贴所选节点（不保留到未选中节点输出的连接）
`Ctrl` + `C`/`Ctrl` + `Shift` + `V`	复制并粘贴所选节点（保留从未选中节点输出到粘贴节点输入的连接）
`Shift` + `Drag`	同时移动多个所选节点
`Ctrl` + `D`	加载默认节点图
`Alt` + `+`	放大画布
`Alt` + `-`	缩小画布
`Ctrl` + `Shift` + LMB + Vertical drag	放大/缩小画布
`P`	固定/取消固定所选节点
`Ctrl` + `G`	编组所选节点
`Q`	显示/隐藏队列
`H`	显示/隐藏历史记录
`R`	刷新节点图
`F`	显示/隐藏菜单
`.`	视图适配选区（未选中时适配整个节点图）
Double-Click LMB	打开节点快速搜索面板
`Shift` + Drag	一次移动多条连接线
`Ctrl` + `Alt` + LMB	断开所点击槽位上的所有连接线

Ctrl 也可以替换为 Cmd，供 macOS 用户使用。

安装

Windows 和 Mac

我们强烈建议使用桌面应用：

Windows 便携版

Windows 提供了一个便携版独立构建包，可用于在 Nvidia GPU 上运行，或仅在 CPU 上运行。不推荐普通用户使用。普通用户应使用上方的桌面应用。

直接下载链接（nvidia）

下载后，使用 7-Zip 或在较新的 Windows 版本中通过 Windows 资源管理器解压并运行。对于较小的模型，通常只需将 checkpoints（巨大的 ckpt/safetensors 文件）放到 ComfyUI\models\checkpoints 中，但许多较大的模型包含多个文件。请务必按照说明操作，以确定应将其放入 ComfyUI\models\ 下的哪个子文件夹。

如果解压时遇到问题，右键单击文件 -> 属性 -> 解除阻止

上述便携版当前内置 python 3.13 和 pytorch cuda 13.0。如果无法启动，请更新你的 Nvidia 驱动程序。

所有官方便携版下载：

适用于 AMD GPU 的便携版

适用于 Intel GPU 的便携版

适用于 Nvidia GPU 的便携版（支持 20 系列及以上）。

包含 pytorch cuda 12.6 和 python 3.12 的 Nvidia GPU 便携版（支持 Nvidia 10 系列及更旧 GPU，请勿在更新的 20 系列及以上 GPU 上使用）。

如何在另一个 UI 和 ComfyUI 之间共享模型？

查看配置文件以设置模型的搜索路径。在 Windows 独立版中，你可以在 ComfyUI 目录中找到此文件。将此文件重命名为 extra_model_paths.yaml，并使用你喜欢的文本编辑器进行编辑。

comfy-cli

你可以使用 comfy-cli 安装并启动 ComfyUI：

pip install comfy-cli
comfy install

手动安装（Windows、Linux）

Python 3.14 可以运行，但部分自定义节点可能存在问题。自由线程版本可以运行，但由于某些依赖项会启用 GIL，因此尚未获得完整支持。

Python 3.13 得到了很好的支持。如果在 3.13 上遇到某些自定义节点依赖项的问题，可以尝试 3.12

torch 2.7 仅做最低支持，强烈建议使用更新版本。Nvidia 20 系列及以上显卡需要使用 cu130 或更高版本的 pytorch。一些功能和优化可能仅在新版本中生效。除最新大版本发布不足两周外，我们通常建议使用 pytorch 最新大版本搭配最新 cuda 版本。如果你的 pytorch 已超过 6 个月，请更新。

安装步骤：

使用 Git 克隆该仓库。

将你的 SD 检查点（较大的 ckpt/safetensors 文件）放入：models/checkpoints

将 VAE 放入：models/vae

AMD GPU（Linux）

AMD 用户如果尚未安装 rocm 和 pytorch，可以通过 pip 安装，安装稳定版本的命令如下：

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm7.2

安装 ROCm 7.2 的 nightly 版本的命令如下，该版本可能包含一些性能提升：

pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/rocm7.2

AMD GPU（实验性支持：Windows 和 Linux），仅支持 RDNA 3、3.5 和 4。

这些版本相比上述构建版本的硬件支持较少，但可在 Windows 上运行。你还需要安装专用于你所用硬件的 pytorch 版本。

RDNA 3（RX 7000 系列）：

pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx110X-all/

RDNA 3.5（Strix halo/Ryzen AI Max+ 365）：

pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx1151/

RDNA 4（RX 9000 系列）：

pip install --pre torch torchvision torchaudio --index-url https://rocm.nightlies.amd.com/v2/gfx120X-all/

Intel GPU（Windows 与 Linux）

Intel Arc GPU 用户可以通过 pip 安装支持 torch.xpu 的 PyTorch 原生版本。更多信息可参见这里

安装 PyTorch xpu，请使用以下命令：

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/xpu

这是安装 PyTorch xpu nightly 的命令，可能会带来一些性能提升：

pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/xpu

NVIDIA

NVIDIA 用户应使用以下命令安装稳定版 PyTorch：

pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu130

如果你想改用 PyTorch nightly，可使用以下命令，可能会有性能提升。

pip install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu132

故障排查

如果遇到 "Torch not compiled with CUDA enabled" 错误，请使用以下命令卸载 torch：

pip uninstall torch

然后使用上面的命令重新安装。

依赖项

在 ComfyUI 文件夹中打开终端，并按如下方式安装依赖项：

pip install -r requirements.txt

完成后，所有组件应该都已安装，可以继续运行 ComfyUI。

其他：

Apple Mac silicon

你可以在搭载 Apple Mac silicon（M1、M2、M3 或 M4）的设备上，使用任意较新的 macOS 版本安装 ComfyUI。

安装 PyTorch nightly。操作说明请参阅 Apple 开发者指南《Mac 上的加速 PyTorch 训练》（请确保安装最新的 PyTorch nightly）。
按照 Windows 和 Linux 的ComfyUI 手动安装说明操作。
安装 ComfyUI 的依赖项。如果你已安装其他 Stable Diffusion UI，可能可以直接复用这些依赖项。
运行 python main.py 启动 ComfyUI

注意：请记得将模型、VAE、LoRA 等添加到对应的 Comfy 文件夹，详见ComfyUI 手动安装。

Ascend NPU

适用于兼容 Ascend PyTorch 扩展（torch_npu）的模型。开始前，请确保你的环境满足安装页面中列出的先决条件。以下是根据你所在平台和安装方式整理的分步指南：

如需要，请先按照 torch-npu 安装页面中的说明，为 Linux 安装推荐版本或更高版本的 kernel。
接着，按照适用于你所在平台的说明安装 Ascend Basekit，其中包括驱动、固件和 CANN。
然后，按照安装页面上对应平台的说明，安装 torch-npu 所需的软件包。
最后，按照 Linux 的ComfyUI 手动安装指南操作。所有组件安装完成后，即可按照前文所述方式运行 ComfyUI。

Cambricon MLU

适用于兼容 Cambricon PyTorch 扩展（torch_mlu）的模型。以下是根据你所在平台和安装方式整理的分步指南：

按照安装页面上对应平台的说明安装 Cambricon CNToolkit
然后，按照安装页面上的说明安装 PyTorch(torch_mlu)
运行 python main.py 启动 ComfyUI

Iluvatar Corex

适用于兼容 Iluvatar PyTorch 扩展的模型。以下是根据你所在平台和安装方式整理的分步指南：

按照安装页面上对应平台的说明安装 Iluvatar Corex Toolkit
运行 python main.py 启动 ComfyUI

ComfyUI-Manager

ComfyUI-Manager 是一款扩展，可方便地为 ComfyUI 安装、更新和管理自定义节点。

安装配置

安装管理器依赖：

pip install -r manager_requirements.txt

运行 ComfyUI 时，通过 --enable-manager 参数启用管理器：
```
python main.py --enable-manager
```

命令行选项

参数	说明
`--enable-manager`	启用 ComfyUI-Manager
`--enable-manager-legacy-ui`	使用旧版管理器 UI 代替新 UI（隐含 `--enable-manager`）
`--disable-manager-ui`	禁用管理器 UI 和端点，同时保留安全检查和定时安装完成等后台功能（需要 `--enable-manager`）

运行

python main.py

针对 ROCm 未官方支持的 AMD 显卡

如果遇到问题，可尝试使用以下命令运行：

适用于 6700、6600 以及其他可能的 RDNA2 或更早显卡： HSA_OVERRIDE_GFX_VERSION=10.3.0 python main.py

适用于 AMD 7600 以及其他可能的 RDNA3 显卡： HSA_OVERRIDE_GFX_VERSION=11.0.0 python main.py

AMD ROCm 提示

可以尝试设置环境变量 PYTORCH_TUNABLEOP_ENABLED=1，它可能会提升运行速度，但首次运行会非常缓慢。

备注

只有输入全部正确并带有输出的图部分才会执行。

只有相邻两次执行之间发生变化的图部分才会执行。如果两次提交相同的图，只有第一次会被执行。如果更改了图的末尾部分，只有被更改的部分以及依赖于它的部分会被执行。

将生成的 png 拖放到网页上，或加载一个 png，即可获得包含创建时使用的随机种子在内的完整工作流。

你可以使用 () 来改变单词或短语的强调权重，例如：(good code:1.2) 或 (bad code:0.8)。() 的默认强调权重为 1.1。若要在实际提示词中使用 () 字符，请通过 \( 或 \) 进行转义。

你可以使用 {day|night} 来编写通配符/动态提示词。使用这种语法时，“{wild|card|test}”会在每次将提示词加入队列时，由前端随机替换为 "wild"、"card" 或 "test"。若要在实际提示词中使用 {} 字符，请通过 \{ 或 \} 进行转义。

动态提示词还支持 C 风格注释，例如 // comment 或 /* comment */。

要在文本提示词中使用文本反转概念/嵌入，请将它们放入 models/embeddings 目录，并在 CLIPTextEncode 节点中按如下方式使用（可省略 .pt 扩展名）：

embedding:embedding_filename.pt

如何显示高质量预览？

使用 --preview-method auto 启用预览。

默认安装包含一种快速的 latent 预览方法，分辨率较低。若要使用 TAESD 启用更高质量的预览，请下载 taesd_decoder.pth, taesdxl_decoder.pth, taesd3_decoder.pth and taef1_decoder.pth 并将其放置到 models/vae_approx 文件夹中。安装完成后，重启 ComfyUI，并使用 --preview-method taesd 启动它，即可启用高质量预览。

如何使用 TLS/SSL？

通过运行以下命令生成自签名证书（不适用于共享/生产用途）和密钥：openssl req -x509 -newkey rsa:4096 -keyout key.pem -out cert.pem -sha256 -days 3650 -nodes -subj "/C=XX/ST=StateName/L=CityName/O=CompanyName/OU=CompanySectionName/CN=CommonNameOrHostname"

使用 --tls-keyfile key.pem --tls-certfile cert.pem 启用 TLS/SSL，此时应用可通过 https://... 访问，而不是 http://...。

注意：Windows 用户可以使用 alexisrolland/docker-openssl 或其中一个第三方二进制发行版来运行上面的命令示例。

如果你使用容器，请注意卷挂载 -v 可以指定相对路径，因此 ... -v ".\:/openssl-certs" ... 会在命令提示符或 PowerShell 终端的当前目录下创建密钥和证书文件。

支持与开发频道

Discord：可以尝试 #help 或 #feedback 频道。

Matrix 空间: #comfyui_space:matrix.org（类似 Discord，但是开源的）。

另请参见：https://www.comfy.org/

悄悄说——我们正在招聘！ 参与构建 ComfyUI：comfy.org/careers

前端开发

自 2024 年 8 月 15 日起，我们已切换到新的前端，目前托管在独立仓库中：ComfyUI Frontend。由 TS/Vue 编译生成的 JS 文件已发布到 pypi，并作为 ComfyUI 的依赖项安装。

报告问题与功能请求

对于任何与前端相关的缺陷、问题或功能请求，请使用 ComfyUI 前端仓库。这有助于我们更高效地管理和解决前端相关问题。

使用最新的前端版本

新版前端现已成为 ComfyUI 的默认前端。但请注意：

主 ComfyUI 仓库中的前端每两周更新一次。
独立的前端仓库中提供每日版本。

若要使用最新的前端版本：

若要使用最新每日版本，请使用以下命令行参数启动 ComfyUI：
```
--front-end-version Comfy-Org/ComfyUI_frontend@latest
```
若要使用指定版本，请将 latest 替换为所需版本号：
```
--front-end-version Comfy-Org/ComfyUI_frontend@1.2.2
```

这种方法可以让你轻松地在稳定的双周版本与前沿的每日更新之间切换，也可以指定特定版本用于测试。

QA

我该选择哪款 GPU 来运行 ComfyUI？

参阅此页面获取一些推荐

项目介绍