Run Local and API LLMs, Features Conditioning manipulation via Omost, supports Ollama, LlamaCPP LMstudio, Koboldcpp, TextGen, Transformers or via APIs Anthropic, Groq, OpenAI, Google Gemini, Mistral, xAI and create your own charcters assistants (SystemPrompts) with custom presets and muchmore
################# ATENTION ####################
It Might comflict with IF_AI_tools so if you have it installed please remove it before installing IF_LLM I am working on adding this tools to IF_AI_tools so you only need one or the other
###############################################
Lighter version of ComfyUI-IF_AI_tools is a set of custom nodes to Run Local and API LLMs and LMMs, supports Ollama, LlamaCPP LMstudio, Koboldcpp, TextGen, Transformers or via APIs Anthropic, Groq, OpenAI, Google Gemini, Mistral, xAI and create your own profiles (SystemPrompts) with custom presets and muchmore
You can technically use any LLM API that you want, but for the best expirience install Ollama and set it up.
To install Ollama models just open CMD or any terminal and type the run command follow by the model name such as
ollama run llama3.2-vision
If you want to use omost
ollama run impactframes/dolphin_llama3_omost
if you need a good smol model
ollama run ollama run llama3.2
Optionally Set enviromnet variables for any of your favourite LLM API keys "XAI_API_KEY", "GOOGLE_API_KEY", "ANTHROPIC_API_KEY", "MISTRAL_API_KEY", "OPENAI_API_KEY" or "GROQ_API_KEY" with those names or otherwise it won't pick it up you can also use .env file to store your keys
[NEW] xAI Grok Vision, Mistral, Google Gemini exp 114, Anthropic 3.5 Haiku, OpenAI 01 preview [NEW] Wildcard System [NEW] Local Models Koboldcpp, TextGen, LlamaCPP, LMstudio, Ollama [NEW] Auto prompts auto generation for Image Prompt Maker runs jobs on batches automatically [NEW] Image generation with IF_PROMPTImaGEN via Dalle3 [NEW] Endpoints xAI, Transformers, [NEW] IF_profiles System Prompts with Reasoning/Reflection/Reward Templates and custom presets [NEW] WF such as GGUF and FluxRedux
ollama run impactframes/llama3_ifai_sd_prompt_mkr_q4km:latest
ollama run impactframes/ifai_promptmkr_dolphin_phi3:latest
https://huggingface.co/impactframes/llama3_if_ai_sdpromptmkr_q4km
https://huggingface.co/impactframes/ifai_promptmkr_dolphin_phi3_gguf
custom_nodes
folder, type CMD
on the address bar to open a command prompt,
and run the following command to clone the repository:
git clone https://github.com/if-ai/ComfyUI-IF_LLM.git
OR
In ComfyUI protable version just dounle click embedded_install.bat
or type CMD
on the address bar on the newly created custom_nodes\ComfyUI-IF_LLM
folder type
H:\ComfyUI_windows_portable\python_embeded\python.exe -m pip install -r requirements.txt
replace C:\
for your Drive letter where you have the ComfyUI_windows_portable directory
On custom environment activate the environment and move to the newly created ComfyUI-IF_LLM
cd ComfyUI-IF_LLM
python -m pip install -r requirements.txt
If you want to use AWQ to save VRAM and up to 3x faster inference you need to install triton and autoawq
pip install triton
pip install --no-deps --no-build-isolation autoawq
I also have precompiled wheels for FA2 sageattention and trton for windows 10 for cu126 and pytorch 2.6.3 and python 12+ https://huggingface.co/impactframes/ComfyUI_desktop_wheels_win_cp12_cu126/tree/main
None yet
ancient Megastructure, small lone figure
You can try out these workflow examples directly in ComfyDeploy!
| Workflow | Try It | |--------------|---------|
|CD_FLUX_LoRA||
|CD_HYVid_I2V_&_T2V_Native_IFLLM||
|CD_HYVid_I2V_&_T2V_i2VLora_Native|
|
|CD_HYVid_I2V_Lora_KjWrapper|
|
If you find this tool useful, please consider supporting my work by: