# Project information site_name: MaaS-Base site_url: https://docs.gpustack.ai site_author: MaaS-Base site_description: >- MaaS-Base is an open-source GPU cluster manager designed for efficient AI model deployment. It lets you run models efficiently on your own GPU hardware by choosing the best inference engines, scheduling GPU resources, analyzing model architectures, and automatically configuring deployment parameters. # Repository repo_name: gpustack/gpustack repo_url: https://github.com/gpustack/gpustack # Copyright copyright: Copyright © 2026 GPUStack.ai theme: name: material logo: assets/logo-white.png favicon: assets/logo.png custom_dir: docs/overrides palette: - media: "(prefers-color-scheme)" toggle: icon: material/brightness-auto name: Switch to light mode - media: "(prefers-color-scheme: light)" scheme: default primary: custom accent: orange toggle: icon: material/brightness-7 name: Switch to dark mode - media: "(prefers-color-scheme: dark)" scheme: slate primary: light blue accent: orange toggle: icon: material/brightness-4 name: Switch to system preference features: - search.suggest - search.highlight - content.tabs.link - navigation.indexes - content.tooltips - navigation.path - navigation.tabs - content.code.annotate - content.code.copy - content.code.select - content.action.view - content.action.edit - navigation.top - navigation.footer - navigation.tracking - search.share - toc.follow icon: repo: fontawesome/brands/github language: en plugins: - glightbox - search - redirects: redirect_maps: 'installation/nvidia/installation.md': 'installation/requirements.md' 'installation/amd/installation.md': 'installation/requirements.md' 'installation/ascend/installation.md': 'installation/requirements.md' 'installation/hygon/installation.md': 'installation/requirements.md' 'installation/mthreads/installation.md': 'installation/requirements.md' 'installation/iluvatar/installation.md': 'installation/requirements.md' 'installation/cambricon/installation.md': 'installation/requirements.md' 'installation/metax/installation.md': 'installation/requirements.md' - minify: minify_html: true minify_css: true minify_js: true cache_safe: true htmlmin_opts: remove_comments: true extra: generator: false version: provider: mike social: - icon: fontawesome/brands/github link: https://github.com/gpustack - icon: fontawesome/brands/docker link: https://hub.docker.com/r/gpustack - icon: fontawesome/brands/python link: https://pypi.org/project/gpustack/ - icon: fontawesome/brands/discord link: https://discord.gg/VXYJzuaqwD - icon: fontawesome/brands/weixin link: https://github.com/gpustack/gpustack/blob/main/docs/assets/wechat-group-qrcode.jpg extra_javascript: - image-selector/assets/extra.js - https://unpkg.com/katex@0/dist/katex.min.js - javascripts/katex.js - https://unpkg.com/katex@0/dist/contrib/auto-render.min.js - https://unpkg.com/tablesort@5.3.0/dist/tablesort.min.js - javascripts/tablesort.js extra_css: - stylesheets/extra.css - https://unpkg.com/katex@0/dist/katex.min.css markdown_extensions: - abbr - def_list - attr_list - md_in_html - pymdownx.highlight: linenums: true line_spans: __span anchor_linenums: true pygments_lang_class: true - pymdownx.inlinehilite - pymdownx.superfences: custom_fences: - name: mermaid class: mermaid format: !!python/name:pymdownx.superfences.fence_code_format - pymdownx.tabbed: alternate_style: true - admonition - sane_lists - pymdownx.details - pymdownx.tilde - pymdownx.betterem - pymdownx.caret - pymdownx.tasklist: custom_checkbox: true - pymdownx.arithmatex: generic: true - pymdownx.emoji: emoji_generator: !!python/name:material.extensions.emoji.to_svg emoji_index: !!python/name:material.extensions.emoji.twemoji - pymdownx.critic - pymdownx.keys - pymdownx.snippets: auto_append: - docs/includes/abbreviations.md edit_uri: edit/main/docs/ nav: - Home: - Overview: overview.md - Quickstart: quickstart.md - Installation: - Requirements: installation/requirements.md - Installation: installation/installation.md - Air-Gapped Installation: installation/air-gapped.md - Uninstallation: installation/uninstallation.md - Upgrade: upgrade.md - Migration: migration.md - User Guide: - Playground: - Playgound: user-guide/playground/index.md - Chat: user-guide/playground/chat.md - Image: user-guide/playground/image.md - Audio: user-guide/playground/audio.md - Embedding: user-guide/playground/embedding.md - Rerank: user-guide/playground/rerank.md - Model Catalog: user-guide/model-catalog.md - Model Deployment Management: user-guide/model-deployment-management.md - Model Route Management: user-guide/model-route-management.md - Model Provider Management: user-guide/model-provider-management.md - Inference Backend Management: user-guide/inference-backend-management.md - Built-in Inference Backends: user-guide/built-in-inference-backends.md - Compatibility Check: user-guide/compatibility-check.md - Model File management: user-guide/model-file-management.md - Cluster Management: user-guide/cluster-management.md - Cloud Credential Management: user-guide/cloud-credential-management.md - API Key Management: user-guide/api-key-management.md - User Management: user-guide/user-management.md - Single Sign-On (SSO) Authentication: user-guide/sso.md - Observability: user-guide/observability.md - Benchmarking: user-guide/benchmarking.md - Using Models: - Using Large Language Models: using-models/using-large-language-models.md - Using Vision Language Models: using-models/using-vision-language-models.md - Using Embedding Models: using-models/using-embedding-models.md - Using Reranker Models: using-models/using-reranker-models.md - Using Image Generation Models: using-models/using-image-generation-models.md - Using Audio Models: using-models/using-audio-models.md - Tutorials: - Running DeepSeek R1 671B with Distributed vLLM: tutorials/running-deepseek-r1-671b-with-distributed-vllm.md - Running DeepSeek R1 671B with Distributed Ascend Mindie: tutorials/running-deepseek-r1-671b-with-distributed-ascend-mindie.md - Inference On CPUs: tutorials/inference-on-cpus.md - Inference with Tool Calling: tutorials/inference-with-tool-calling.md - Using Custom Inference Backend: tutorials/using-custom-backends.md - Adding a GPU Cluster Using DigitalOcean: tutorials/adding-gpucluster-using-digitalocean.md - Adding a GPU Cluster Using Kubernetes: tutorials/adding-gpucluster-using-kubernetes.md - Managing Model Routes: tutorials/managing-model-routes.md - Integrations: - Inference APIs: integrations/inference-apis.md - Integrate with Claude Code: integrations/integrate-with-claude-code.md - Integrate with Dify: integrations/integrate-with-dify.md - Integrate with RAGFlow: integrations/integrate-with-ragflow.md - Integrate with CherryStudio: integrations/integrate-with-cherrystudio.md - Integrate with OpenClaw: integrations/integrate-with-openclaw.md - Integrate with n8n: integrations/integrate-with-n8n.md - Integrate with MaxKB: integrations/integrate-with-maxkb.md - Architecture: architecture.md - Scheduler: scheduler.md - Troubleshooting: troubleshooting.md - FAQ: faq.md - API Reference: api-reference.md - CLI Reference: - Start: cli-reference/start.md - Download Tools: cli-reference/download-tools.md - Reload Config: cli-reference/reload-config.md - List Images: cli-reference/list-images.md - Save Images: cli-reference/save-images.md - Load Images: cli-reference/load-images.md - Copy Images: cli-reference/copy-images.md - Environment Variables: environment-variables.md - Inference Performance Lab: - Overview: performance-lab/overview.md - Optimizing Throughput: - Qwen3.5-35B-A3B: - H200: performance-lab/qwen3.5-35b-a3b/h200.md - Qwen3.5-9B: - H100: performance-lab/qwen3.5-9b/h100.md - GLM-4.5-Air: - A100: performance-lab/glm-4.5-air/a100.md - H100: performance-lab/glm-4.5-air/h100.md - GLM-4.x: - A100: performance-lab/glm-4.x/a100.md - H100: performance-lab/glm-4.x/h100.md - H200: performance-lab/glm-4.x/h200.md - GPT-OSS-20B: - A100: performance-lab/gpt-oss-20b/a100.md - H100: performance-lab/gpt-oss-20b/h100.md - GPT-OSS-120B: - A100: performance-lab/gpt-oss-120b/a100.md - H100: performance-lab/gpt-oss-120b/h100.md - DeepSeek-R1: - H200: performance-lab/deepseek-r1/h200.md - DeepSeek-V3.2: - H200: performance-lab/deepseek-v3.2/h200.md - Qwen3-8B: - 910B: performance-lab/qwen3-8b/910b.md - Qwen3-14B: - A100: performance-lab/qwen3-14b/a100.md - H100: performance-lab/qwen3-14b/h100.md - Qwen3-32B: - A100: performance-lab/qwen3-32b/a100.md - H100: performance-lab/qwen3-32b/h100.md - Qwen3-30B-A3B: - 910B: performance-lab/qwen3-30b-a3b/910b.md - Qwen3-235B-A22B: - A100: performance-lab/qwen3-235b-a22b/a100.md - H100: performance-lab/qwen3-235b-a22b/h100.md - Optimizing Latency: - Qwen3.5-35B-A3B: - H200: performance-lab/qwen3.5-35b-a3b/h200-latency.md - Qwen3.5-9B: - H100: performance-lab/qwen3.5-9b/h100-latency.md - Qwen3-8B: - H100: performance-lab/qwen3-8b/h100-latency.md - References: - The Impact of Quantization on vLLM Inference Performance: performance-lab/references/the-impact-of-quantization-on-vllm-inference-performance.md - Evaluating LMCache Prefill Acceleration in vLLM: performance-lab/references/evaluating-lmcache-prefill-acceleration-in-vllm.md - Container Image Selector: image-selector/index.html