Back to search
100
MCP live MCP 2025-11-25 streamable-http

FitLLM

run.fitllm/fitllm

Will this LLM fit on your GPU, multi-GPU rig or Mac? Exact VRAM & KV-cache math. Read-only.

Uptime
100.0%
1 direct probes · 30d
Response
788ms
last probe
Tools
3
callable
Resources
4
readable
Prompts
0
available
Primary URL
https://fitllm.run/api/mcp

Tools · 3

check_llm_fit

Check whether a specific local LLM fits in the memory of a specific GPU or Apple Silicon Mac. Returns fits/tight/won't-fit verdict with the full memory breakdown (weights, KV cache, overhead), max con…

what_fits_on_hardware

Rank which popular local LLMs fit on a given GPU or Apple Silicon Mac (at ~4-bit quantization, 8K context) — models that fit come first, biggest first, with max context each. Use when a user asks "wha…

list_supported

List the built-in model names and hardware names this fit-checker knows (for mapping user wording to exact names). Standard text-only HuggingFace transformer configs can also be checked via fitllm.run…

Resources · 4

models

Built-in local LLM catalog with the architecture fields the fit engine uses (layers, KV heads/dims, sliding window, MLA, hybrid attention, MoE). JSON.

fitllm://models
hardware

GPUs (VRAM, bandwidth, source-cited) and Apple Silicon chips (unified memory tiers) known to the engine. JSON.

fitllm://hardware
census

Every model × device × quant verdict computed by the engine — starter matrix inline; full CSV/JSON at fitllm.run/data. CC0 data.

fitllm://census
engine

The open MIT calculation engine behind every verdict: source, npm package, conformance vectors, versions. Agents may fetch and run it directly.

fitllm://engine

How to use

Add to your Claude Desktop / Cursor / Cline MCP config:

{
  "mcpServers": {
    "fitllm": {
      "url": "https://fitllm.run/api/mcp",
      "transport": "streamable-http"
    }
  }
}