RigBase is an MCP server published by com.getrigbase, hosted at getrigbase.com. When last checked on 2026-11-22 it was working: it answers and lists its tools. It declares 4 tools, including search_builds, can_it_run, best_gpu_for_model, get_build.
Measured local-LLM tok/s on home AI rigs. Search builds, can-it-run, best GPU; sourced + linked.
1 connections through 5 shared values. Tap any node.
Inferred from the host name and headers: what a server runs on, not who runs it.
Search RigBase measured home AI builds. Filter by GPU name, model name, minimum VRAM (GB), and/or maximum build cost. Returns top matches with tok/s, quant, engine, source credit and canonical getrigbase.com URL. Never invents numbers.
| Parameter | Type | Description |
|---|---|---|
| gpu | string | Substring to match in the GPUs field (e.g. "3090", "DGX Spark", "5090"). |
| limit | integer | Max builds to return (default 8, max 25). |
| model | string | Substring to match a measured model name (e.g. "Qwen3.8", "gpt-oss 120B"). |
| currency | string | Optional currency filter: EUR or USD. |
| max_price | number | Maximum reported build cost (same currency as the build, usually EUR or USD). |
| min_vram_gb |
None yet. Changes appear after a server's second daily check.
Squares are this server's values; dots are other servers sharing them. A shared address or publisher is a lead, not proof of a common operator.
| number |
| Minimum total VRAM / unified memory in GB. |
{
"name": "search_builds",
"description": "Search RigBase measured home AI builds. Filter by GPU name, model name, minimum VRAM (GB), and/or maximum build cost. Returns top matches with tok/s, quant, engine, source credit and canonical getrigbase.com URL. Never invents numbers.",
"inputSchema": {
"type": "object",
"properties": {
"gpu": {
"type": "string",
"description": "Substring to match in the GPUs field (e.g. \"3090\", \"DGX Spark\", \"5090\")."
},
"limit": {
"type": "integer",
"description": "Max builds to return (default 8, max 25)."
},
"model": {
"type": "string",
"description": "Substring to match a measured model name (e.g. \"Qwen3.8\", \"gpt-oss 120B\")."
},
"currency": {
"type": "string",
"description": "Optional currency filter: EUR or USD."
},
"max_price": {
"type": "number",
"description": "Maximum reported build cost (same currency as the build, usually EUR or USD)."
},
"min_vram_gb": {
"type": "number",
"description": "Minimum total VRAM / unified memory in GB."
}
}
},
"annotations": {
"readOnlyHint": true,
"openWorldHint": false
}
}Ask whether a given GPU setup can run a model. Returns measured RigBase results when available; otherwise a VRAM-fit estimate clearly labelled as an estimate (weights arithmetic, not a measured speed). Advises against quants below Q4. Always includes source credit and URLs for measured rows.
| Parameter | Type | Description |
|---|---|---|
| gpu * | string | GPU or system name, e.g. "RTX 3090", "RTX 5090", "DGX Spark", "Apple Silicon Mac". |
| count | integer | How many identical GPUs / nodes (default 1). |
| model * | string | Model name, e.g. "Qwen3.8 27B", "gpt-oss 120B", "Gemma 4 31B". |
| quant | string | Optional quant for the estimate (default Q4_K_M). Avoid below Q4. |
{
"name": "can_it_run",
"description": "Ask whether a given GPU setup can run a model. Returns measured RigBase results when available; otherwise a VRAM-fit estimate clearly labelled as an estimate (weights arithmetic, not a measured speed). Advises against quants below Q4. Always includes source credit and URLs for measured rows.",
"inputSchema": {
"type": "object",
"required": [
"gpu",
"model"
],
"properties": {
"gpu": {
"type": "string",
"description": "GPU or system name, e.g. \"RTX 3090\", \"RTX 5090\", \"DGX Spark\", \"Apple Silicon Mac\"."
},
"count": {
"type": "integer",
"description": "How many identical GPUs / nodes (default 1)."
},
"model": {
"type": "string",
"description": "Model name, e.g. \"Qwen3.8 27B\", \"gpt-oss 120B\", \"Gemma 4 31B\"."
},
"quant": {
"type": "string",
"description": "Optional quant for the estimate (default Q4_K_M). Avoid below Q4."
}
}
},
"annotations": {
"readOnlyHint": true,
"openWorldHint": false
}
}Return the fastest measured RigBase result for a model, plus tok/s per euro where a build cost or known GPU street price exists. Includes source credit and canonical URL.
| Parameter | Type | Description |
|---|---|---|
| limit | integer | How many top results to return (default 5, max 15). |
| model * | string | Model name or substring, e.g. "Gemma 4 31B" or "Qwen3.8". |
{
"name": "best_gpu_for_model",
"description": "Return the fastest measured RigBase result for a model, plus tok/s per euro where a build cost or known GPU street price exists. Includes source credit and canonical URL.",
"inputSchema": {
"type": "object",
"required": [
"model"
],
"properties": {
"limit": {
"type": "integer",
"description": "How many top results to return (default 5, max 15)."
},
"model": {
"type": "string",
"description": "Model name or substring, e.g. \"Gemma 4 31B\" or \"Qwen3.8\"."
}
}
},
"annotations": {
"readOnlyHint": true,
"openWorldHint": false
}
}Fetch one RigBase build by slug (e.g. "dual-3090-ryzen-3700x"). Returns parts, measured results, source credit and canonical URL.
| Parameter | Type | Description |
|---|---|---|
| slug * | string | Build slug from a RigBase /builds/.../ URL. |
{
"name": "get_build",
"description": "Fetch one RigBase build by slug (e.g. \"dual-3090-ryzen-3700x\"). Returns parts, measured results, source credit and canonical URL.",
"inputSchema": {
"type": "object",
"required": [
"slug"
],
"properties": {
"slug": {
"type": "string",
"description": "Build slug from a RigBase /builds/.../ URL."
}
}
},
"annotations": {
"readOnlyHint": true,
"openWorldHint": false
}
}RigBase is a database of measured local-LLM speeds (tokens/sec) on real home AI rigs (NVIDIA/AMD/Intel GPUs, Apple Silicon, Strix Halo, DGX Spark). Every measured number links to its public original source with credit (e.g. Mia's AI Lab, Tech2wild). Use search_builds / get_build / best_gpu_for_model for measured data; use can_it_run for measured matches first, and only then a labelled VRAM-fit estimate. Always cite the source credit and the canonical https://getrigbase.com URL from the tool result. Never invent tok/s numbers. Advise against quants below Q4.
Names, descriptions and schemas were written by the server and are shown as data.