Under Validation

Stop hand-tuning n_gpu_layers until the OOM.

A local MCP advisor for developers already running GGUF models. Profiles hardware and suggests (or applies) offload, context, and batch settings for llama.cpp and Ollama. Not a new inference engine.

Join the waitlist See how it works

Concept validation only. No payment, no commitment, no recurring newsletter.

Decision brief Validating

Why an MCP autotuner?

Local LLM runners still leave developers iterating n_gpu_layers, context length, and batch size to avoid OOM while keeping tok/s usable. Guides and Reddit threads restate the same trial-and-error loop. This concept is a thin host profiler exposed over Model Context Protocol so an agent in Cursor or Claude Desktop can query safe settings. Free scripts and LM Studio UIs already help many people — this waitlist tests residual paid demand.

Best fit: Developers running local GGUF models via llama.cpp or Ollama who already use MCP-capable coding agents.
Not for: Cloud-only inference users, people who only use fully automatic GUI apps and never touch flags, or teams needing a new model runtime.
Current status: Concept validation; no production build yet

Local-first directionPortable outputsBuilt around a specific workflowEarly access is free

Proposed workflow

What the tuner would do

Advisor first. Hot-reload of every backend parameter is not assumed; some runners still need a restart.

Host-aware recommendations

Read RAM/VRAM/thermals and model size/quant hints; propose n_gpu_layers, context, and batch envelopes.

MCP tool surface

Expose tune/recommend tools to agents already in the developer workflow instead of a separate dashboard.

Local only

No telemetry of prompts or model weights. Runs beside your existing Ollama or llama.cpp install.

The honest status

This is a product decision, not a finished product.

The concept is being validated before development time is committed. Joining tells us the problem is relevant to you and gives you first access if the evidence supports a build.

You will not be charged.
Your email is only used for this experiment.
You can leave the list at any time.

Before you decide

Questions, answered clearly.

The current scope, privacy model, and next step without launch-day promises.

Does this replace Ollama or LM Studio?

No. It sits beside them. If your GUI already picks safe defaults, you may not need it.

Will it hot-reload Ollama without restart?

Not guaranteed. Some parameters require a model reload. The product hypothesis includes honest limits in the UX.

Does this guarantee the intended outcome?

No. The concept should state its limits clearly and preserve human review wherever judgment is required.

What happens when I request access?

Your email is recorded for this experiment only. You receive one relevant update or beta invitation if the concept moves forward.

Early access

Want a local MCP inference autotuner?

Join the waitlist. Planned pricing around a €19 one-time license or low monthly if demand clears the bar — no charge today.