Canary-ComfyUI

Canary-ComfyUI requires a minimum of 1536MB (1.5GB) of VRAM for base operation. For optimal performance, a GPU with at least 6GB of VRAM is recommended. This node supports low VRAM mode for resource-constrained setups.

✅ RTX 3060 (12GB): Yes, fully compatible with 9.3GB headroom. ✅ RTX 4070 (12GB): Yes, fully compatible with 9.3GB headroom. ✅ RTX 4070 Ti (16GB): Yes, fully compatible with 12.9GB headroom. ✅ RTX 4090 (24GB): Yes, fully compatible with 20.1GB headroom

On an RTX 3060 (12GB VRAM), Canary-ComfyUI runs smoothly on an RTX 3060 (12GB) with 9.3GB of headroom. This is sufficient to run the node alongside standard SD 1.5 and SDXL workflows in full precision. On an RTX 4090 (24GB VRAM), the node runs with extreme headroom on an RTX 4090 (24GB) with 20.1GB of dedicated headroom. This allows you to combine the node with massive models (like FLUX.1 Dev, Schnell, or Hunyuan Video) in full precision (FP16) without any offload flags.

Canary-ComfyUI requires the following PyTorch-related packages: pytorch-lightning==2.5.4, pytorch-metric-learning==2.9.0, torch-audiomentations==0.12.0, torch-pitch-shift==1.2.5, torchmetrics==1.8.1. Ensure your ComfyUI environment has these installed. Ensure your PyTorch installation matches your CUDA version (use torch.version.cuda to check).

To run Canary-ComfyUI, you need to install: absl-py==2.3.1, accelerate==1.10.1, aiofiles==24.1.0, aiohappyeyeballs==2.6.1, aiohttp==3.12.15, aiosignal==1.4.0, alembic==1.16.5, annotated-types==0.7.0, antlr4-python3-runtime==4.9.3, anyio==4.10.0, asteroid-filterbanks==0.4.0, asttokens==3.0.0, async-timeout==5.0.1, attrs==25.3.0, audioread==3.0.1, backports-datetime-fromisoformat==2.0.3, braceexpand==0.1.7, brotli==1.1.0, certifi==2025.8.3, cffi==1.17.1, charset-normalizer==3.4.3, click==8.2.1, cloudpickle==3.1.1, colorama==0.4.6, coloredlogs==15.0.1, colorlog==6.9.0, contourpy==1.3.2, cycler==0.12.1, cytoolz==1.0.1, datasets==4.0.0, decorator==5.2.1, dill==0.3.8, docopt==0.6.2, editdistance==0.8.1, einops==0.8.1, exceptiongroup==1.3.0, executing==2.2.0, fastapi==0.116.1, ffmpy==0.6.1, fiddle==0.3.0, filelock==3.13.1, flatbuffers==25.2.10, fonttools==4.59.2, frozenlist==1.7.0, fsspec==2024.6.1, future==1.0.0, gitdb==4.0.12, gitpython==3.1.45, gradio==5.44.1, gradio-client==1.12.1, graphviz==0.21, greenlet==3.2.4, groovy==0.1.2, grpcio==1.74.0, h11==0.16.0, httpcore==1.0.9, httpx==0.28.1, huggingface-hub==0.34.4, humanfriendly==10.0, hydra-core==1.3.2, hyperpyyaml==1.2.2, idna==3.10, indic-numtowords==1.1.0, inflect==7.5.0, intervaltree==3.1.0, ipython==8.37.0, jedi==0.19.2, jinja2==3.1.4, jiwer==3.1.0, joblib==1.5.2, julius==0.2.7, kaldi-python-io==1.2.2, kiwisolver==1.4.9, lazy-loader==0.4, levenshtein==0.27.1, lhotse==1.30.3, libcst==1.8.2, librosa==0.11.0, lightning==2.4.0, lightning-utilities==0.15.2, lilcom==1.8.1, llvmlite==0.44.0, loguru==0.7.3, mako==1.3.10, markdown==3.8.2, markdown-it-py==4.0.0, markupsafe==2.1.5, marshmallow==4.0.1, matplotlib==3.10.6, matplotlib-inline==0.1.7, mdurl==0.1.2, mediapy==1.1.6, ml-dtypes==0.5.3, more-itertools==10.7.0, mpmath==1.3.0, msgpack==1.1.1, multidict==6.6.4, multiprocess==0.70.16, nemo-toolkit==2.4.0, networkx==3.3, num2words==0.5.14, numba==0.61.2, numpy==1.26.4, omegaconf==2.3.0, onnx==1.19.0, onnxruntime==1.22.1, optuna==4.5.0, orjson==3.11.3, packaging==24.2, pandas==2.3.2, parso==0.8.5, peft==0.17.1, pillow==11.3.0, pip==25.2, piper-tts==1.3.0, plac==1.4.5, platformdirs==4.4.0, pooch==1.8.2, primepy==1.3, prompt-toolkit==3.0.52, propcache==0.3.2, protobuf==5.29.5, psutil==7.0.0, pure-eval==0.2.3, pyannote-audio==3.3.2, pyannote-core==5.0.0, pyannote-database==5.1.3, pyannote-metrics==3.2.1, pyannote-pipeline==3.0.1, pyarrow==21.0.0, pybind11==3.0.1, pycparser==2.22, pydantic==2.11.7, pydantic-core==2.33.2, pydub==0.25.1, pygments==2.19.2, pyloudnorm==0.1.1, pyparsing==3.2.3, pyreadline3==3.5.4, python-dateutil==2.9.0.post0, python-multipart==0.0.20, pytorch-lightning==2.5.4, pytorch-metric-learning==2.9.0, pytz==2025.2, pyyaml==6.0.2, rapidfuzz==3.14.0, regex==2025.8.29, requests==2.32.5, resampy==0.4.3, rich==14.1.0, ruamel-yaml==0.18.15, ruamel-yaml-clib==0.2.12, ruff==0.12.11, sacremoses==0.1.1, safehttpx==0.1.6, safetensors==0.6.2, scikit-learn==1.7.1, scipy==1.15.3, semantic-version==2.10.0, semver==3.0.4, sentencepiece==0.2.1, sentry-sdk==2.35.1, setuptools==80.9.0, shellingham==1.5.4, six==1.17.0, smmap==5.0.2, sniffio==1.3.1, sortedcontainers==2.4.0, soundfile==0.13.1, sox==1.5.0, soxr==0.5.0.post1, speechbrain==1.0.3, sqlalchemy==2.0.43, stack-data==0.6.3, starlette==0.47.3, sympy==1.13.3, tabulate==0.9.0, tensorboard==2.20.0, tensorboard-data-server==0.7.2, tensorboardx==2.6.4, termcolor==3.1.0, text-unidecode==1.3, texterrors==1.0.9, threadpoolctl==3.6.0, tokenizers==0.21.4, tomli==2.2.1, tomlkit==0.13.3, toolz==1.0.0, torch-audiomentations==0.12.0, torch-pitch-shift==1.2.5, torchmetrics==1.8.1, tqdm==4.67.1, traitlets==5.14.3, transformers==4.51.3, typeguard==4.4.4, typer==0.17.3, typing-extensions==4.12.2, typing-inspection==0.4.1, tzdata==2025.2, urllib3==2.5.0, uv==0.8.14, uvicorn==0.35.0, wandb==0.21.3, wcwidth==0.2.13, webdataset==1.0.2, websockets==15.0.1, werkzeug==3.1.3, wget==3.2, whisper-normalizer==0.1.12, win32-setctime==1.2.0, wrapt==1.17.3, xxhash==3.5.0, yarl==1.20.1. You can install these using pip or add them to your requirements.txt file.

Canary-ComfyUI supports low VRAM mode. To reduce memory usage: (1) Enable --lowvram or --medvram flags in ComfyUI, (2) Reduce batch size to 1, (3) Use fp16 or fp8 precision if supported, (4) Close other GPU applications.

To install Canary-ComfyUI: (1) Navigate to your ComfyUI/custom_nodes directory, (2) Clone the repository: git clone https://github.com/Juste-Leo2/Canary-ComfyUI, (3) Install dependencies: pip install -r requirements.txt (if present), (4) Restart ComfyUI. Alternatively, use ComfyUI Manager for one-click installation.