I will set up or fix comfyui and turn your workflow into a private api


About this gig
ComfyUI throwing red nodes, import errors, out-of-memory crashes, or just not starting? I get it installed, fixed and running on your own machine, then, if you need it, turn your workflow into a private API your app can call.
You receive a working setup, a written health report of what was wrong and what I changed, a test image from your workflow, and notes so you can keep it running.
I can help with:
- Fresh installs on Windows with NVIDIA, Mac with Apple Silicon, Linux, or a cloud GPU
- Fixing missing custom nodes, failed imports, broken updates and Python or CUDA errors
- Getting SDXL, Flux and other workflows running, with the right models in the right folders
- Speed and VRAM tuning for your GPU
- Your workflow as an HTTP API with token access, input limits and docs
I use my own diagnostic script that checks the Python environment, GPU, custom nodes and every model a workflow needs, so fixes start from facts, not guesswork.
Before ordering, send your OS, GPU, the error or a screenshot, and the workflow file if you have one.
Unsure which package fits? Message me first and I will confirm or send a custom offer.
Get to know Nishant
AI Developer for Local LLM, RAG, Agents and Automation
- FromIndia
- Member sinceAug 2018
- Last delivery1 year
Languages
English, Spanish, Portuguese
My Portfolio
FAQ
Do you need remote access to my computer?
Not always. Many fixes work by you running my diagnostic script and sharing the report. If hands-on access is needed, it happens after the order is placed, in a session you watch and end yourself. I never ask for your passwords.
Will it work on my GPU?
NVIDIA cards with enough VRAM for your models work best; Apple Silicon Macs work and are slower. Tell me your GPU and VRAM before ordering and I will say honestly what it can run, including low VRAM options.
Are the models included?
Models are downloaded from their official sources during setup, under their own licenses. Some, such as certain Flux versions, restrict commercial use. I point out the license so you can choose the right one.
Where does the private API run?
On your own GPU machine or on a cloud GPU account you own, such as RunPod. You keep the account, the keys and any hosting costs. The API only exposes the inputs you approve.
Can you build a new workflow from scratch?
Yes, as a custom offer after we agree the goal and models. These packages cover setup, repair, getting an existing or template workflow running, and turning it into an API.
What will you not set up?
Workflows for explicit content, deepfakes of real people without consent, or anything that breaks a model license or Fiverr's terms. If unsure, ask first.

