Search
Every page on Agent Infra Guide, grouped by type. To filter as you type, open search from the header or press ⌘ K — Ctrl K on Windows and Linux.
Most people search for the hardware. Describe the job instead — the infrastructure follows from it.
Try “Best AI receptionist for a dental office” →Hubs
- Hub/Self-Hosted AI Agent Infrastructure — Agent Infra GuideIndependent infrastructure guides for running self-hosted AI agents 24/7 on managed platforms, VPS infrastructure, GPU cloud, and local hardware.
- Hub/agentsAI Agent RuntimesCompare persistent AI agent runtimes for local hosting. Open-source, self-hosted, no cloud dependency.
- Hub/agent-toolsAI Infrastructure API & Agent Tools — Agent Infra GuideMachine-readable model requirements, hardware records and compatibility evidence for AI agents, developer tools and infrastructure workflows.
- Hub/hostingAI Agent Hosting: Compare VPS, Managed, Dedicated & GPU CloudChoose AI agent hosting based on workload, uptime, control and cost. Compare VPS, managed platforms, dedicated servers, GPU cloud and local hardware, then use free sizing and cost tools.
Comparisons
- Comparison/hosting/best-ai-agent-hostingBest AI Agent Hosting: Managed Platforms, VPS and Self-HostingCompare AI agent hosting options including VPS, managed platforms, GPU cloud and local hardware. Learn how to host an AI agent for reliable 24/7 operation.
- Comparison/hosting/best-vps-ai-agentsBest VPS for AI Agents in 2026: RAM, CPU & Deployment GuideChoose a VPS for AI agents by RAM, CPU, storage and workload, then deploy a persistent 24/7 agent with secure access, restart policy, monitoring and backups.
- Comparison/hosting/self-hosted-ai-agentsBest Self-Hosted AI Agents: Frameworks, Hardware and Maintenance ComparedCompare self-hosted AI-agent approaches by deployment model, infrastructure requirements, maintenance burden, security, and operational control.
- Comparison/gpu-cloud/providersGPU Cloud Providers Compared: Pricing, VRAM & LLM Fit (2026)Compare GPU cloud providers for LLM training and inference by VRAM, GPU rental pricing, persistence, availability, storage and operational fit.
- Comparison/gpu-cloud/qwenBest GPU Cloud for Qwen: GPU, VRAM and Provider RecommendationsChoose GPU cloud infrastructure for Qwen models by parameter count, quantization, VRAM, context length, persistence, and workload type.
- Comparison/local-ai/best-macBest Mac for Local LLMs: Unified-Memory Tiers and Model FitCompare current Macs for local LLM inference by unified-memory capacity, memory bandwidth, model size, quantization, and workload requirements.
- Comparison/local-ai/ollama-hostingBest Ollama Hosting: Managed GPU, VPS and Self-Hosted OptionsCompare ways to run Ollama through local hardware, VPS infrastructure, GPU cloud, and Ollama cloud models for persistent AI workloads.
- Comparison/local-ai/lm-studio-vs-ollamaLM Studio vs Ollama 2026: Local LLM Server ComparisonCompare LM Studio vs Ollama for local LLMs, APIs, model management, headless servers, network access, agents and private AI infrastructure.
- Comparison/mcp/hostingMCP Server Hosting: Best Platforms for Secure Remote MCPPlan secure remote MCP server infrastructure with HTTP transport, authentication, scaling, observability, deployment, and authorization controls.
- Comparison/agent-ops/observability-toolsBest AI Agent Observability Tools: Tracing, Evals and Runtime MonitoringCompare the capabilities needed for AI-agent observability, including traces, tool calls, logs, metrics, evaluations, failures, latency, and cost.
- Comparison/agent-ops/security-toolsBest AI Agent Security Tools: Identity, Sandboxing, MCP and Runtime ControlsUnderstand the security controls needed for AI agents, including least privilege, identity, sandboxing, tool authorization, MCP security, audit logs, and approvals.
- Comparison/coding-agents/best-ai-coding-agentsBest AI Coding Agents: Compare Task Fit, Cost, Autonomy and SecurityCompare AI coding agents by repository access, autonomy, sandboxing, parallel work, testing, approvals, security controls, and real software-engineering tasks.
- Comparison/local-ai/mac-miniMac Mini for Local AI: M6 and M5 Pro Unified-Memory GuideCompare M6 and M5 Pro Mac mini configurations for local AI. Understand how unified memory, bandwidth, and model size determine which Mac mini to buy.
- Comparison/local-ai/local-vs-cloud-gpuLocal AI vs Cloud GPU: Cost, Privacy and Uptime ComparedCompare local AI hardware with cloud GPU infrastructure by cost, privacy, utilization, scaling, uptime and control.
- Comparison/agents/openclaw/hostingOpenClaw Hosting: Best VPS, Managed, Docker & Local OptionsCompare OpenClaw hosting options across VPS, managed platforms, Docker, local hardware and cloud infrastructure, with requirements, security and operational tradeoffs.
- Comparison/agents/hermes-vs-openclawHermes Agent vs OpenClaw: Hosting, Models, Security and AutomationCompare Hermes Agent and OpenClaw from an infrastructure perspective: hosting, remote access, persistence, gateways, deployment, and operational control.
- Comparison/agents/openclaw/alternativesOpenClaw Alternatives: Self-Hosted AI Agent Frameworks ComparedCompare OpenClaw with Hermes Agent and other self-hosted AI agent frameworks by runtime, deployment model, messaging support, local models, security, and infrastructure requirements.
Guides
- Guide/tools/guidy-ai-reviewGuidy AI Review: Pricing, Screen Guidance & Who It's ForAn evidence-based Guidy AI review: free vs $9.99/month, Windows and Mac support, privacy caveats, refund terms, and when screen guidance beats chat or agents.
- Guide/benchmarksMeasured Local LLM Benchmarks: Speed and MemoryOllama benchmarks on an Apple M4 Pro with 24 GiB unified memory: decode tokens/s, sampled peak process memory, task outcomes, and downloadable measurement data.
- Guide/products/agent-infrastructure-recommendation-kitAgent Infrastructure Recommendation Kit — Build a Defensible AI Infrastructure PlanA six-step professional workflow for sizing AI workloads, comparing local, VPS, GPU cloud and hybrid infrastructure, documenting evidence, and delivering a defensible recommendation.
- Guide/hosting/ai-dedicated-serverDedicated AI Server: GPU Providers, Prices & When You Need OneDedicated AI server costs and GPU rentals explained. Compare bare-metal servers vs VPS and hourly GPU cloud, calculate monthly break-even, and inspect dated provider prices.
- Guide/guides/run-ai-agent-24-7How to Run an AI Agent 24/7: Hosting, State, Queues and RecoveryA practical guide to infrastructure for a persistent AI agent: hosting, state, queues, supervision, recovery, logs, backups, and security.
- Guide/models/qwen25-coder/hardware-requirementsQwen2.5-Coder Hardware Requirements (2026): 7B, 14B & 32B RAM / VRAMQwen2.5-Coder RAM and VRAM requirements from 0.5B to 32B. Loaded at 4K context: 7B ~4.9 GB, 14B ~9.8 GB, 32B ~21 GB — and why 32B on a 24 GB GPU is short-context only.
- Guide/models/llama32/hardware-requirementsLlama 3.2 Hardware Requirements (2026): 1B & 3B RAM / VRAMLlama 3.2 1B and 3B RAM and VRAM requirements by quantization and context. 3B needs ~2.5 GB at 4K context — but about 17 GB at the full 128K, because the KV cache outgrows the model.
- Guide/models/phi4/hardware-requirementsPhi-4 Hardware Requirements (2026): RAM & VRAM for 14BPhi-4 14B RAM and VRAM requirements by quantization and context. 4-bit weight floor ~7.3 GB, Ollama Q4_K_M artifact 9.1 GB, about 12.5 GB loaded at the full 16K context.
- Guide/models/qwen/hardware-requirementsQwen Hardware Requirements (2026): RAM & VRAM by ModelQwen minimum system requirements by model: RAM, VRAM, published Ollama file sizes, 8/16/24 GB sizing examples, and measured Qwen2.5 results.
- Guide/models/deepseek/hardware-requirementsDeepSeek Hardware Requirements: RAM and VRAM by Model SizeEstimate DeepSeek RAM and VRAM requirements for R1 distills and the full R1 model using parameter count, quantization, and explicit runtime headroom.
- Guide/models/llama/hardware-requirementsLlama Hardware Requirements: RAM and VRAM by Model SizeEstimate Llama RAM and VRAM requirements for Llama 3.2 and Llama 3.1 models using official parameter counts and AIG's shared memory-sizing formula.
- Guide/models/gemma/hardware-requirementsGemma Hardware Requirements: RAM and VRAM for Gemma 3Estimate Gemma 3 RAM and VRAM requirements for 1B, 4B, 12B and 27B models using official parameter counts and transparent quantization math.
- Guide/models/mistral/hardware-requirementsMistral Hardware Requirements: RAM and VRAM for 7B and 24B ModelsEstimate Mistral RAM and VRAM requirements for Mistral 7B and Mistral Small 24B using official model cards and AIG's shared sizing formula.
- Guide/models/gpt-oss/hardware-requirementsgpt-oss Hardware Requirements: RAM and VRAM for 20B and 120BEstimate gpt-oss RAM and VRAM requirements using OpenAI's 21B and 117B total parameter counts, with explicit notes for MoE active parameters.
- Guide/agent-ops/jevJev for AI Agents: System One Model & PricingWhat Jev is, what it costs, how to access it, and where a typed-decision model fits in an AI agent's control layer: routing, tool gating, loop and completion checks, trace grading, and output checks.
- Guide/local-ai/serverLocal AI Server: Hardware, VRAM, Cost & Build Guide (2026)Build a local AI server for LLMs and agents. Size RAM and VRAM, compare GPU and Apple silicon hardware, and plan storage, networking, security and cost.
- Guide/local-ai/ram-requirementsLocal AI RAM Requirements: Model Size, Context and ConcurrencyPlan RAM and VRAM for local AI using model size, quantization, context length and concurrency.
- Guide/guides/how-much-vram-llmHow Much VRAM Do I Need for an LLM? A Practical Sizing GuideEstimate LLM VRAM from model weights, quantization, context, runtime overhead and concurrency.
- Guide/guides/ai-agent-server-requirementsAI Agent Server Requirements: CPU, RAM, GPU, Storage and UptimePlan server requirements for persistent AI agents across CPU, RAM, GPU, storage, networking, recovery and 24/7 operation.
- Guide/agents/openclaw/securityOpenClaw Security 2026: Is OpenClaw Safe? Hardening ChecklistSecure OpenClaw with gateway authentication, pairing, allowlists, sandboxing, tool restrictions, private networking and the built-in security audit.
- Guide/agents/openclaw/vps-setupBest VPS for OpenClaw? Requirements, Providers & SetupCompare VPS options for OpenClaw, documented requirements, provider candidates, Docker or native setup, backups, remote access and gateway security.
- Guide/agents/hermes/best-vpsBest VPS for Hermes Agent: Requirements & Test CandidatesCompare Hermes Agent VPS requirements, hosting options and test candidates. AIG has not named a best VPS yet because RAM, CPU and recovery measurements are still pending.
- Guide/agents/openclaw/mac-miniOpenClaw on Mac Mini: Setup, Local Models and Always-On HostingRun OpenClaw on a Mac mini with Apple Silicon. Compare cloud-model, local-model, and VPS deployment architectures with current Mac mini configurations and pricing.
- Guide/agents/openclaw/costOpenClaw Cost: Software, Hosting, API and Total Infrastructure PricingBreak down the cost of running OpenClaw: software, VPS hosting, model API usage, local inference hardware, and optional services.
Tools
- Tool/tools/dedicated-server-advisorDedicated AI Server Advisor: GPU VRAM, RAM and Hosting CostsEstimate GPU count, host RAM, CPU and storage for dedicated LLM inference servers. Compare dedicated and cloud quotes with explicit memory and workload assumptions.
- Tool/tools/agent-hosting-cost-calculatorAI Agent Hosting Cost CalculatorEstimate monthly AI agent hosting costs across VPS, GPU cloud, managed platforms, and local hardware using your own current prices.
- Tool/tools/ai-token-cost-calculatorAI Token Cost Calculator: Estimate LLM API CostsEstimate AI API costs per request, month, and year. Calculate LLM token spending using input, output, cached-token pricing, and request volume.
- Tool/tools/openclaw-security-checkerOpenClaw Security Checker 2026: Audit Your ConfigurationCheck a sanitized OpenClaw JSON configuration for public gateway exposure, disabled authentication, sandboxing, broad tool permissions and possible secrets.
- Tool/tools/local-vs-cloud-ai-costLocal AI vs GPU Cloud Cost Calculator 2026Compare local AI hardware cost against GPU cloud cost using your own hardware price, power, hourly GPU rate and usage. Share the exact comparison.
- Tool/tools/infrastructure-planner-liteFree AI Infrastructure Planner Lite: Memory, Hosting & ArchitecturePlan AI agent infrastructure with a free interactive RAM/VRAM estimate and local, VPS, GPU cloud or hybrid decision screen. Transparent assumptions, no account required.
- Tool/tools/llm-vram-calculatorLLM RAM and VRAM CalculatorEstimate LLM RAM and VRAM from model size, quantization, runtime overhead, and context reserve with a transparent planning formula.
- Tool/tools/can-i-run-this-modelCan I Run This AI Model? Hardware Compatibility CheckerCheck whether Qwen, Gemma or Llama models are likely to fit Mac unified memory or GPU VRAM using current model artifact evidence and explicit runtime headroom.
- Tool/tools/architecture-finderAI Agent Architecture Finder: Local, VPS, GPU Cloud or HybridAnswer five questions about inference, data, uptime, memory and budget to get a deterministic local, VPS, GPU cloud or hybrid recommendation with its reasoning.