How to run Flowise with a self-hosted model on a cloud GPU
Rent a GPU on Deploy, then open a terminal in JupyterLab or over SSH. Install Flowise with the npm or Docker commands the README prints, serve a model on that machine, and point ChatOllama or ChatLocalAI at the base URL that page prints. Copy the chatflow and the host directory the compose comment names before you destroy the instance, because destroying the instance deletes its disk.
An agent can rent the GPU and run this application on it. The agent pages are GPU compute for AI agents and the agent integration guide. Those pages are how an agent discovers a GPU and runs work on it. This page is the Flowise install. A Docker Compose app whose self-host page opens http://localhost/install is How to run Dify with a self-hosted model on a cloud GPU. A flow editor whose Ollama component defaults the base URL to http://127.0.0.1:11434 is How to run Langflow with a self-hosted model on a cloud GPU. A CLI whose connect help sets --base-url to http://localhost:8000/v1 for openai-compatible is How to run Letta agents with a self-hosted model on a cloud GPU. A browser chat whose serve command listens on port 8080 is How to run Open WebUI on a cloud GPU. A Python agent library whose CodeAgent can take TransformersModel or an OpenAI-compatible server is How to run smolagents with a self-hosted model on a cloud GPU. A crew whose LLM class takes base_url for Ollama or an OpenAI-compatible server is How to run CrewAI with a self-hosted model on a cloud GPU. A graph runtime whose overview example does not name a model is How to run LangGraph agents with a self-hosted model on a cloud GPU. An AgentChat library whose Ollama example sets model to llama3.2 is How to run AutoGen agents with a self-hosted model on a cloud GPU. A typed agent whose vLLM example sets base_url to http://localhost:8000/v1 is How to run Pydantic AI agents with a self-hosted model on a cloud GPU. A framework whose README sets Ollama and HuggingFaceEmbedding is How to build LlamaIndex agents with a self-hosted model on a cloud GPU. A browser agent whose models page sets base_url to http://localhost:11434/v1 is How to run browser-use with a self-hosted model on a cloud GPU. A pipeline whose OllamaChatGenerator sets url to http://localhost:11434 is How to run Haystack agents with a self-hosted model on a cloud GPU. A program whose language-models page sets dspy.LM to ollama_chat/llama3.2:1b is How to run DSPy with a self-hosted model on a cloud GPU. An agent whose Ollama example sets id to llama3.1:8b is How to run Agno agents with a self-hosted model on a cloud GPU. A TypeScript agent whose local-models section sets url on the model object is How to run Mastra agents with a self-hosted model on a cloud GPU. A ChatAgent whose vLLM example sets ModelPlatformType.VLLM is How to run CAMEL-AI agents with a self-hosted model on a cloud GPU. A general vLLM install is How to serve an OpenAI-compatible API with vLLM on a cloud GPU. A general SGLang install is How to run SGLang on a cloud GPU. Ollama’s install and port 11434 are How to run Ollama on a cloud GPU. An embedding server is How to serve embeddings and rerankers with Text Embeddings Inference on a cloud GPU. CloudGPU’s hosted gateway is a different base URL: How do I use an OpenAI-compatible API for open models?. How published weight sizes are read: How much VRAM do I need to run an LLM?. The tunnel is How to use SSH port forwarding on a cloud GPU. If the process runs out of memory: How to fix CUDA out of memory on a cloud GPU. If the process raises no kernel image on an RTX 5090: How to fix CUDA no kernel image on an RTX 5090.
The README says source code in this repository is made available under the Apache License Version 2.0, and LICENSE.md says content under packages/server/src/enterprise is under a separate commercial license.
The commands below are copied from the upstream files and docs pages named in each step. They were not run on a CloudGPU instance for this page. Anything this page adds around them is marked. The docs index lists other chat-model pages. This page does not copy those names.
What the docs print about memory
The Flowise README, the ChatOllama page, the ChatLocalAI page, and the Agentflow V2 page print no GPU GB figure. The ChatLocalAI page says LocalAI runs on consumer grade hardware. It does not print a GB figure. The model information table on unsloth/Llama-3.1-8B-Instruct prints parameter counts 8B, 70B, and 405B, and a context length of 128k. This page does not treat those parameter counts as vram_gb. A later section of that card names training hardware. This page does not copy that section and does not use it as an inference memory figure. This page does not mark a catalog row as enough.
GET /v1/public/compute-capabilities does not list GPUs. It returns feature flags. The memory column is vram_gb from GET /v1/public/gpu-catalog (generated_at 2026-11-14T09:10:32.863Z). The catalog response also includes an availability field. This page does not copy it. Every catalog row below has min_gpu_count 1. Which listing you can start is on Deploy.
The catalog template list had no Flowise template. It did have Ollama templates (jupyterlab-ollama, ollama-qwen3, ollama). Those templates are a different install. This page does not treat them as Flowise.
| Catalog name | vram_gb |
|---|---|
| RTX 2080 | 8 |
| RTX 3080 | 10 |
| RTX 2080 Ti | 11 |
| RTX 3080 Ti | 12 |
| V100 16G | 16 |
| A30 | 24 |
| RTX 3090 | 24 |
| RTX 4090 | 24 |
| RTX 4090D | 24 |
| RTX 4080 SUPER 32G | 32 |
| RTX 5090 | 32 |
| RTX 5090D | 32 |
| V100 32G | 32 |
| A100 40G | 40 |
| A800 40G | 40 |
| A40 | 48 |
| RTX 4090 48G | 48 |
| RTX 4090D 48G | 48 |
| RTX A6000 | 48 |
| Ascend 910B | 64 |
| NVIDIA RTX PRO 5000 72G | 72 |
| RTX PRO 5000 72G | 72 |
| A100 80G | 80 |
| A800 80G | 80 |
| H800 | 80 |
| RTX 6000D | 83 |
| H20 96G | 96 |
| RTX PRO 6000 96G | 96 |
| RTX PRO 6000 Max-Q | 96 |
| H20-3e 141G | 141 |
Install
Main of FlowiseAI/Flowise on 2026-11-14 was commit 9291856d1ea4a4ceea9f8fef8ce14f4f6c81e8eb, committed 2026-08-13T12:38:19Z. pushed_at was the same timestamp. The GitHub license field was NOASSERTION. The README at that commit has this heading: “Flowise has been archived.” It points at discussion 6727. That discussion’s title is “The Future of Flowise”, created 2026-08-13T12:33:57Z. This page did not read the discussion body.
Release tag flowise@3.1.4 is commit a65f81bb43ef66d3ce734bf0dff4223ae8041c95, committed 2026-07-29T09:59:33Z, published 2026-07-29T15:51:25Z. The README at that tag does not contain the archive heading. Main and that tag are not the same commit. package.json at both commits sets version to 3.1.4 and engines to node ^24 and pnpm ^10.26.0. npm flowise latest on 2026-11-14 was 3.1.4, license SEE LICENSE IN LICENSE.md, engines node ^24, published 2026-07-29T10:15:33.837Z. The README quick start says NodeJS >= 20.0.0. That sentence and the engines field are not the same string. This page did not install the package.
The quick start, copied from the main README:
npm install -g flowise
npx flowise start
That README says to open http://localhost:3000.
The Docker Compose steps on the same README are: clone the project, go to the docker folder, copy .env.example to .env, run docker compose up -d, open http://localhost:3000, and stop with docker compose stop. docker/.env.example at the main commit sets PORT=3000. docker/docker-compose.yml publishes that PORT value and its healthcheck requests http://localhost: plus that port plus /api/v1/ping. With the example file, that URL is http://localhost:3000/api/v1/ping. The healthcheck also sets start_period to 30s. This page does not treat that field as a startup time.
The Docker image lines, copied:
docker build --no-cache -t flowise .
docker run -d --name flowise -p 3000:3000 flowise
docker stop flowise
The developer section also prints pnpm install, pnpm build, and pnpm start, still on http://localhost:3000, and says pnpm dev reloads on http://localhost:8080. Port 8080 is that developer path. It is not the quick start port. The README links other deployment hosts. This page does not copy that list. The files this page read do not print an SSH command. For a server you start on the instance, the tunnel is How to use SSH port forwarding on a cloud GPU.
Agent and chatflow
The ChatOllama page, fetched 2026-11-14, says Chat Models > drag ChatOllama, then fill in the model that is running. Its example model name is llama2. Its Ollama Docker lines, copied:
docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama
docker exec -it ollama ollama run llama3
The same page’s “Running on Docker” section says that if both Flowise and Ollama are on Docker, change the ChatOllama Base URL. It prints http://host.docker.internal:8000 for Windows and MacOS, and http://172.17.0.1:8000 for Linux. Those URLs use port 8000. The Docker run line above publishes port 11434. Those ports are not the same. This page does not pick one for you.
The Agentflow V2 page documents an Agent node. Its Model parameter says it specifies the model from a chosen service. The example on that line names vendors. This page does not copy those names. The page also documents an LLM node with the same kind of example. This page does not copy that example either.
ChatLocalAI and an OpenAI-compatible base path
The ChatLocalAI page says LocalAI is a drop-in replacement REST API that’s compatible with the OpenAI API specifications. Its setup lines clone https://github.com/go-skynet/LocalAI, enter that directory, and copy a model file into models/. The example download command on that page names a file this page does not copy. Then:
docker compose up -d --pull always
The page says the API is then accessible at localhost:8080, and it prints:
curl http://localhost:8080/v1/models
The sample JSON on that page names a model file. This page does not copy that id. The Flowise fields it prints are Base Path http://localhost:8080/v1 and Model Name equal to a file inside /models. If both Flowise and LocalAI are on Docker, it prints http://host.docker.internal:8080/v1, and for Linux http://172.17.0.1:8080/v1.
The ChatOllama and ChatLocalAI pages do not print hf download. The block below is added on this page. How to download a Hugging Face model prints pip install -U "huggingface_hub" and hf download. This page did not download the repository. The Flowise files this page read do not print this repository id.
pip install -U "huggingface_hub"
hf download unsloth/Llama-3.1-8B-Instruct
On 2026-11-14, GET https://huggingface.co/api/models/unsloth/Llama-3.1-8B-Instruct reported gated false, cardData.license llama3.1, pipeline_tag text-generation, lastModified 2025-05-09T19:32:57.000Z, and sha 4699cc75b550f9c6f3173fb80f4703b62d946aa5. The card YAML also includes extra_gated_prompt. Those two fields are not the same fact. This page did not run hf download.
How to serve an OpenAI-compatible API with vLLM on a cloud GPU is the general server install. The Flowise pages this page read do not name vLLM. How to run Ollama on a cloud GPU is the Ollama install.
When it fails
CUDA out of memory. The Flowise files and docs pages this page read print no GPU GB figure. If the model server raises CUDA out of memory, use how to fix CUDA out of memory on a cloud GPU. This page does not pick a smaller model id for you.
“CUDA error: no kernel image is available for execution on the device” on an RTX 5090. The Flowise README does not name compute capability 12.0. The RTX 5090 guide says that card is compute capability 12.0 (sm_120) and that PyTorch 2.7 and later, on a build for CUDA 12.8 or newer, is the line that announced support. This page did not install a torch wheel and did not re-test this program on an RTX 5090. Catalog RTX 5090 and RTX 5090D are 32. That is the catalog number only.
What to copy off, then destroy
The compose file comment names the host directory ~/.flowise. Copy that directory, and export the chatflow from the UI, before you destroy the instance. Save your work before you destroy a cloud GPU is the checklist. The terms say data on local instance storage is permanently deleted when the instance is destroyed.
The billing page states the billing rule in these words:
Per-minute options have a 1-minute minimum; unused prepaid minutes are refunded when you destroy the instance. Some configurations use whole-hour billing with a 1-hour minimum; unused time within a charged hour is not refunded on voluntary stop. The billing period, minimum charge and initial prepayment are shown before confirmation. No subscription.
Amounts are not copied here.
Frequently asked questions
What memory figure do the Flowise docs print?
The README, the ChatOllama page, the ChatLocalAI page, and the Agentflow V2 page print no GPU GB figure. The unsloth/Llama-3.1-8B-Instruct model information table prints parameter counts 8B, 70B, and 405B, and context length 128k. This page does not treat those parameter counts as vram_gb. This page does not mark a catalog row as enough. Which listing you can start is on https://cloudgpu.app/deploy.
Which install lines do the docs print?
The README prints npm install -g flowise, then npx flowise start, then http://localhost:3000. It also prints docker compose up -d from the docker folder, and docker run -d --name flowise -p 3000:3000 flowise. docker/.env.example sets PORT=3000. The README says NodeJS >= 20.0.0. package.json engines and the npm engines field say node ^24. Those strings are not the same. npm flowise on 2026-11-14 was 3.1.4. Main README says the repository has been archived. The tag flowise@3.1.4 README does not. This page did not install the package.
How do the docs point Flowise at a local or OpenAI-compatible server?
The ChatOllama page prints docker run with port 11434 and ollama run llama3, and the model example llama2. Its both-on-Docker base URLs use port 8000: http://host.docker.internal:8000 and http://172.17.0.1:8000. Those ports are not the same. The ChatLocalAI page says the API is compatible with the OpenAI API specifications, prints Base Path http://localhost:8080/v1, and curl http://localhost:8080/v1/models. The pages this page read do not name vLLM. The files do not print an SSH command.
What license does the file print?
The README says source code in this repository is made available under the Apache License Version 2.0, and LICENSE.md says content under packages/server/src/enterprise is under a separate commercial license. The GitHub license field on 2026-11-14 was NOASSERTION. npm license for flowise 3.1.4 was SEE LICENSE IN LICENSE.md. That license text is not a CloudGPU rate.
Sources (read 2026-11-14, Asia/Shanghai)
- FlowiseAI/Flowise README at commit
9291856d1ea4a4ceea9f8fef8ce14f4f6c81e8eb, committed 2026-08-13T12:38:19Z. Archive heading points at discussion 6727. Discussion title “The Future of Flowise”, created 2026-08-13T12:33:57Z. This page did not read the discussion body. GitHub license fieldNOASSERTION.package.jsonversion3.1.4, engines node^24and pnpm^10.26.0. Tagflowise@3.1.4is commita65f81bb43ef66d3ce734bf0dff4223ae8041c95, committed 2026-07-29T09:59:33Z. That tag README does not contain the archive heading. Itspackage.jsonversion and engines match main. npmflowise3.1.4, published 2026-07-29T10:15:33.837Z, licenseSEE LICENSE IN LICENSE.md, engines node^24 - LICENSE.md at main: content outside
packages/server/src/enterpriseis Apache License Version 2.0. The copyright line is Copyright © 2023-present FlowiseAI, Inc. The enterprise directory is under a commercial license. The README’s license section says Apache License Version 2.0 docker/.env.exampleat main setsPORT=3000.docker/docker-compose.ymlpublishes that port and healthcheckshttp://localhost:plus the port plus/api/v1/ping.start_periodis30s. This page does not treat that field as a startup time- ChatOllama https://docs.flowiseai.com/integrations/langchain/chat-models/chatollama, fetched 2026-11-14. Model example
llama2. Docker publish11434. Both-on-Docker base URLs use port8000 - ChatLocalAI https://docs.flowiseai.com/integrations/langchain/chat-models/chatlocalai, fetched 2026-11-14. Base Path
http://localhost:8080/v1. Linux Docker base pathhttp://172.17.0.1:8080/v1. The example download filename is not copied - Agentflow V2 https://docs.flowiseai.com/using-flowise/agentflowv2, fetched 2026-11-14. Agent node Model parameter. The vendor example on that line is not copied
- Hugging Face model API
unsloth/Llama-3.1-8B-Instruct:gatedfalse,cardData.licensellama3.1,pipeline_tagtext-generation,lastModified2025-05-09T19:32:57.000Z,sha4699cc75b550f9c6f3173fb80f4703b62d946aa5. The card YAML includesextra_gated_prompt. The model information table prints 8B, 70B, 405B, and 128k.hf downloadis added on this page and was not run. The training-hardware section is not copied - Public GPU memory: https://cloudgpu.app/v1/public/gpu-catalog (
generated_at2026-11-14T09:10:32.863Z).GET /v1/public/compute-capabilitiesreturned feature flags and no GPU list. The catalog had no Flowise template
Cards you can start are on the Deploy page. Amounts are not copied here.
Last updated: 2026-11-14 (Asia/Shanghai).