Determines if open-weight LLMs fit on specific GPUs and provides hardware recommendations and VRAM estimates.
Also known as nodegrove/vram-mcp, vram-mcp, @nodegrove/vram-mcp
- Stars
- 0
- Tools
- 4
- Downloads, 30 days
- —
- Latest version
- 1.1.5
- Last commit
- Nov 3, 2026
- License
- MIT
What is Nodegrove VRAM MCP?
Nodegrove VRAM determines if open-weight LLMs fit on specific GPUs and provides hardware recommendations and VRAM estimates. It exposes 4 tools, such as can_i_run, what_fits, estimate_vram.
Summary generated from the project's README
About Nodegrove VRAM MCP server
Nodegrove VRAM is an MCP server designed for developers and AI enthusiasts who need to determine if specific open-weight large language models can run on their hardware. It provides precise VRAM estimates by calculating the combined requirements of model weights, KV cache, and overhead. The server uses actual model configuration files and hardware specifications rather than benchmarks or guesses to ensure accuracy.
Users can connect to this server remotely via a hosted URL or run it locally using Node.js and stdio. It works with various AI clients including Claude, Cursor, VS Code, and LM Studio. No accounts or API keys are required to use the service. The server offers specialized tools like can_i_run to check compatibility for specific context lengths and what_fits to find the best models for a given GPU. It also includes estimate_vram and estimate_from_hf_repo to analyze any Hugging Face repository by reading its config.json file directly.
What you can do with it
- 1Check model hardware compatibilityDetermine if a specific model like Llama 3.3 70B will fit on an RTX 4090 with a set context length.
- 2Find optimal models for hardwareIdentify the largest or best performing models that can run on a specific graphics card.
- 3Estimate VRAM for specific settingsGet a breakdown of weights and KV cache requirements for a model at a specific quantization level.
Written from the project's README. Every use case is checked against a quote from it.
How to install Nodegrove VRAM MCP
Formats verified 10 Nov 2026Add to CursorAdd to GooseInstall Nodegrove VRAM MCP in Claude Code
$ claude mcp add vram-mcp -- npx -y @nodegrove/vram-mcpInstall Nodegrove VRAM MCP in Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"vram-mcp": {
"command": "npx",
"args": [
"-y",
"@nodegrove/vram-mcp"
]
}
}
}Install Nodegrove VRAM MCP in VS Code
.vscode/mcp.json
{
"servers": {
"vram-mcp": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@nodegrove/vram-mcp"
]
}
}
}Install Nodegrove VRAM MCP in VS Code CLI
$ code --add-mcp '{"name":"vram-mcp","type":"stdio","command":"npx","args":["-y","@nodegrove/vram-mcp"]}'Install Nodegrove VRAM MCP in Claude Desktop
claude_desktop_config.json
{
"mcpServers": {
"vram-mcp": {
"command": "npx",
"args": [
"-y",
"@nodegrove/vram-mcp"
]
}
}
}Install Nodegrove VRAM MCP in Codex
$ codex mcp add vram-mcp -- npx -y @nodegrove/vram-mcpInstall Nodegrove VRAM MCP in Codex config.toml
~/.codex/config.toml
[mcp_servers.vram-mcp] command = "npx" args = ["-y", "@nodegrove/vram-mcp"]
Install Nodegrove VRAM MCP in Gemini CLI
$ gemini mcp add -s user vram-mcp npx -- -y @nodegrove/vram-mcpInstall Nodegrove VRAM MCP in Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"vram-mcp": {
"command": "npx",
"args": [
"-y",
"@nodegrove/vram-mcp"
]
}
}
}Nodegrove VRAM MCP tools 4
Detected from the READMEcan_i_runwhat_fitsestimate_vramestimate_from_hf_repo
| can_i_run | — | |
| what_fits | — | |
| estimate_vram | — | |
| estimate_from_hf_repo | — |
Automated checks
Ran Nov 13, 2026Full report →Passed our automated checks. That is not a security audit: read the code and give the server the narrowest access it needs. How checks work
- Source repository linked and reachablenodegrove/vram-mcp is public and reachablegithub.com/nodegrove/vram-mcp ↗Pass
- Package names this repository as its source@nodegrove/vram-mcp names nodegrove/vram-mcp as sourcenpmjs.com/package/@nodegrove/vram-mcp ↗Pass
- Build provenance (Sigstore attestation)No build attestation published for this releaseNot checked
- Open source license (OSI)MITspdx.org/licenses/MIT.html ↗Pass
- Commit or release in the last yearLast commit or release 2026-11-03github.com/nodegrove/vram-mcp ↗Pass
- OpenSSF ScorecardNot checked
- Package name not close to a popular packageNo popular package one edit away from @nodegrove/vram-mcpPass
- No critical advisory in the package or its dependencies4 resolved dependencies checked; no critical advisoryosv.dev ↗Pass
- No hidden instructions to the model4 tool descriptions and the README read; no override phrases, credential paths or hidden charactersPass
- Destructive tools are labelledNo tool named like delete, drop or execPass
- Remote endpoint answers the MCP handshakeinitialize answered with protocol 2025-06-18 in 205 msPass
README
View on GitHubCan my RTX 4090 run Llama 3.3 70B?
No: Llama 3.3 70B at Q4_K_M with 8,192 tokens of context needs 45.8 GB, and the RTX 4090 holds 22.8 GB after headroom. Short by 23 GB. What would work instead:
- No context length helps: the weights alone are 43.1 GB before a single token of conversation.
- RTX 6000 Ada (48 GB) is within a whisker: 45.8 GB against 45.6 GB after headroom. With the KV cache at Q8 it needs 44.4 GB and fits.
- The smallest card here that runs it exactly as asked: A100, 80 GB usable.
- The biggest model your card does run at these settings, counting mixture-of-experts models at their dense equivalent: Gemma 4 31B, 20.9 GB at ~39 tokens/s.
That is the server's own answer, word for word, from the same code as the GPU compatibility checker on nodegrove.io. Every figure is a stated formula over the model's config.json and the card's published specs: no benchmarks, no guesses. It is free, read-only, and needs no account or key.
Connect
Remote: nothing to install
https://mcp.nodegrove.io/mcp
# Claude Code (add --scope user to have it in every project)
claude mcp add --transport http nodegrove-vram https://mcp.nodegrove.io/mcp
# Codex CLI
codex mcp add nodegrove-vram --url https://mcp.nodegrove.io/mcp
# Gemini CLI
gemini mcp add -s user --transport http nodegrove-vram https://mcp.nodegrove.io/mcp
- Claude (claude.ai and the desktop app): Customize → Connectors → Add → Add custom connector. Name it
Nodegrove VRAM, paste the URL, choose No sign-in, then turn it on in a chat from + → Connectors. - ChatGPT: Settings → Security and login → turn on Developer mode. At chatgpt.com/plugins, add the URL with No Authentication, then pick it in a chat from + → Developer mode.
Most other clients take this in their mcp.json. Cline, Zed, Open WebUI, Devin Desktop and the rest are covered step by step at nodegrove.io/mcp.
{ "mcpServers": { "nodegrove-vram": { "url": "https://mcp.nodegrove.io/mcp" } } }
Local: over stdio
Needs Node.js 20 or newer:
{
"mcpServers": {
"nodegrove-vram": {
"command": "npx",
"args": ["-y", "@nodegrove/vram-mcp"]
}
}
}
Tools
| Tool | Ask it | It answers with |
|---|---|---|
can_i_run |
Can my RTX 4090 run Llama 3.3 70B with 32k of context? | Fits, tight or no; the memory split; a speed ceiling; the longest context that fits; and on a no, every change that would make it fit |
what_fits |
What is the best model for my 16 GB card? | Every model checked on one card, with a recommended everyday pick, the largest that fits, the best at Q8 and the first out of reach |
estimate_vram |
How much VRAM does Qwen3 32B need at 64k? | Weights, KV cache and overhead at each quantisation, and the smallest common card that holds each |
estimate_from_hf_repo |
How much memory does Qwen/Qwen3-Next-80B-A3B-Instruct need? |
Any Hugging Face repo, read from its config.json: the attention layout, the cost of each 1,000 tokens of context, and memory at every quantisation |
list_models, list_gpus |
Which GPUs do you know? | The models and cards with their specs, ids and pages |
Models can be named the way people type them ("llama 3.3 70b", Llama-3.3-70B-Instruct) or given as any Hugging Face repo id. Cards can be named ("4090", "M4 Max") or described by their memory and bandwidth. Every answer carries the numbers as fields, links to the model and card pages on nodegrove.io, and the assumptions behind each figure. All six tools are read-only.
How the numbers are made
- Memory is weights + KV cache + overhead. Weights are parameters × bytes per parameter at the quantisation: FP16 2.00, Q8_0 1.06, Q6_K 0.82, Q5_K_M 0.71, Q4_K_M 0.58, Q3_K_M 0.47. Overhead is 0.5 GB plus 4% of the weights.
- The KV cache is counted the way each model caches. Sliding-window layers stop at their window, hybrid models (linear attention or Mamba) grow a cache only on their few full-attention layers, and latent attention stores one compressed vector per layer. A standard-transformer formula would overstate these models several times over at long context.
- Fit means at most 95% of the memory a runtime can address; above 85% it is tight. Apple silicon gives the GPU about 75% of unified memory.
- Speed is 0.7 × memory bandwidth ÷ bytes of active weights read per token. It is a single-stream ceiling, not a measurement, and the faster the figure, the further real runtimes fall below it.
- The model table was read from each model's
config.jsonand checked against Hugging Face. Any other repo is read live by the same rules, which reproduce every row of the table; anything the reader cannot model is named in the answer, never guessed.
The full method, with every constant, is at nodegrove.io/data. The same figures are an open dataset under CC BY 4.0, DOI 10.5281/zenodo.22966137.
The math as a library
The server is a thin layer over @nodegrove/llm-math, the package nodegrove.io's pages and calculators are built on:
npm install @nodegrove/llm-math
import { modelById, estimate } from '@nodegrove/llm-math';
estimate({ ...modelById('llama-3.3-70b')!, quant: 'q4', context: 8192 }).totalGb; // 45.77
Privacy
The remote server runs on Cloudflare Workers. Nodegrove keeps no request logs and no record of what you ask: each request is answered by a fresh, stateless instance and forgotten. Cloudflare keeps standard edge logs for a short period, as for any website. Requests are rate-limited to 120 a minute per address. When you ask about a Hugging Face repo, the server fetches that repo's public config.json and metadata; only the repo name is sent. The local version contacts nothing but huggingface.co, and only when you ask about a repo there. Details: nodegrove.io/privacy.
Run your own
The remote endpoint is mcp/src/worker.ts. To deploy a copy to your Cloudflare account, change the route in mcp/wrangler.jsonc, then:
cd mcp && npm install && npx wrangler deploy
Development
cd llm-math && npm install && npm test
cd mcp && npm install && npm test && npm run build
llm-math/ is the package of formulas and data; mcp/ is the server, with the stdio entry (src/stdio.ts), the Worker (src/worker.ts) and the tools (src/tools.ts). This repository is published from Nodegrove's main repository, where the site uses the same package. Issues and pull requests are welcome here; accepted changes are applied upstream and credited.
Found a figure that disagrees with its source? Open an issue with the model or card and the link. To report a security problem, see SECURITY.md.
Licence
The code is MIT. The model and GPU data, llm-math/src/models.ts and llm-math/src/gpus.ts, is CC BY 4.0: use it for anything, and credit Nodegrove (nodegrove.io).
Model and GPU names are trademarks of their owners.
Nodegrove VRAM MCP questions
How do I add Nodegrove VRAM MCP to Claude Code?
Run claude mcp add vram-mcp -- npx -y @nodegrove/vram-mcp in your terminal. Add --scope user to make it available in every project.
How do I add Nodegrove VRAM MCP to Cursor?
Open ~/.cursor/mcp.json and add the server under mcpServers, using the Cursor config shown in the install section above.
What tools does Nodegrove VRAM MCP provide?
It exposes 4 tools, including can_i_run, what_fits, estimate_vram, estimate_from_hf_repo.
How does the server calculate VRAM requirements?
It calculates memory as the sum of weights, KV cache, and overhead, using specific formulas for different model architectures like sliding-window or hybrid models.
Is the server free to use?
Yes, the server is free, read-only, and does not require an account or an API key.