# Project Maya

> Open-source engine, server and dashboard for running GLM-5.3-Flash and DeepSeek-V4.1-Flash-1017 locally on one or more GPUs with OpenAI- and Anthropic-compatible APIs.

Project Maya is an open-source local inference project created and led by Peasant Smith (@mw00) and co-maintained with @needmorevram. It runs very large mixture-of-experts models, GLM-5.3-Flash (321B parameters) and DeepSeek-V4.1-Flash-1017 (748B parameters), on hardware users already own, from one NVIDIA GPU up to 16, with experimental AMD and Windows support. The README states that nothing leaves the user's machine.

## What It Is

Maya is a self-hosted inference stack made up of a model engine, a server, a browser dashboard, and quantization tools. Each supported model has its own engine. The engine keeps the most-used experts on the GPU, the next ones in RAM, and the rest on an NVMe SSD, moving them as a conversation proceeds. It measures the GPUs, CPU, RAM and SSD and tunes itself to them.

## How You Use It

A setup script asks which model and quant to download, recommends one for the machine, builds the engine, and starts the server at a local address. The browser interface offers chat, a live hardware and model monitor, image attachments, and thinking levels (Off, Low, High, Max). Apps and coding agents can connect through an OpenAI-compatible API, an Anthropic-compatible messages endpoint (usable with Claude Code), and a Responses API with streaming, function tool calls and schema-validated output. Up to 8 requests can be served at once for GLM-5.3-Flash.

## Models and Quants

The project publishes its own quants of GLM-5.3-Flash (Maya-S, S24, M, M-Derisked, L) with measured accuracy against the FP8 model, plus its own Maya-S quant of DeepSeek-V4.1-Flash-1017. Community GGUF and EXL3 quants can also be used. Optional features include DFlash2 drafting, KV streaming for long contexts, and a calibration step that tunes speed to the machine. The README lists user-reported benchmark numbers across many GPU configurations.

## Requirements and Setup Path

The documented requirements are an NVIDIA GPU with compute capability 7.0 or newer (or an experimental AMD GPU with ROCm 7), 32 GB RAM, a fast NVMe SSD with room for the model, and Linux with CUDA 12.x, g++ and Python 3.10+. Windows 10/11 is experimental and WSL2 is not supported.

## Features
- Runs GLM-5.3-Flash (321B) and DeepSeek-V4.1-Flash-1017 (748B) locally
- Tiered expert placement across GPU VRAM, RAM and NVMe SSD
- Supports one to 16 NVIDIA GPUs, experimental AMD and Windows
- OpenAI-compatible API and Anthropic-compatible messages endpoint
- Responses API with streaming, tool calls and schema-validated output
- Browser chat and live hardware monitor
- Image input support
- Thinking levels: Off, Low, High, Max
- Up to 8 parallel requests for GLM-5.3-Flash
- DFlash2 drafting and KV streaming options
- Self-tuning calibration to local hardware
- Project-made quants plus community GGUF and EXL3 options
- One-command guided setup and report tool

## Integrations
Claude Code, OpenAI-compatible clients, Anthropic-compatible clients, Hugging Face, llama.cpp (ggml), exllamav3, Strata

## Platforms
WINDOWS, LINUX, WEB, API, VSC_EXTENSION, CLI

## Pricing
Open Source

## Version
v1.1.27

## Links
- Website: https://huggingface.co/peasantsmith
- Repository: https://github.com/mw00/project-maya
- EveryDev.ai: https://www.everydev.ai/tools/project-maya