macrostack
Browse

The AI stack

Categories

Local & Sovereign AINotes & KnowledgeObservability & MonitoringPassword ManagersWeb AnalyticsTeam ChatSmart HomeNetworking & RoutersVideo ConferencingCloud Storage & SyncPhotos & MediaAPI DevelopmentImage EditingWorkflow Automation & iPaaSDeveloper Tools & ContainersOffice & Productivity SuitesNo-Code DatabasesCode Hosting & Git ForgesProject ManagementEmail Marketing & NewslettersScheduling & BookingError Tracking & Exception MonitoringLog Management & SIEMVPN & PrivacyEmail & Secure MailVector Databases & AI SearchLLM & Agent FrameworksDomains & Web HostingData Removal & PrivacyAuthentication & IdentityHelp Desk & Customer SupportCloud & VPSKubernetes & Container PlatformsEmbedding ModelsPDF & DocumentsAI Coding AssistantsAI Voice & SpeechLLM Observability & EvaluationLLM Gateways & RoutingCloud GPU & AI ComputeCI/CD & build automationData & pipeline orchestrationModel serving & inferenceAI agent frameworksBackend as a serviceSecrets managementFeature flags & experimentationProduct analyticsSearch infrastructureUptime & status monitoringAffiliate & partner platformsVisitor identification & personalisationWikis & internal docsIdentity & access managementData warehouses & analytics enginesCustomer data platformsCRMObject storageBI & dashboardsE-signatureWhiteboards & diagrammingIn-memory data stores & cachingPlatform as a serviceTransactional & bulk emailHeadless CMSDesign & prototypingE-commerce platformsInternal tools & admin panelsManaged databasesForms & surveysFine-Tuning & Model TrainingRAG & Retrieval PlatformsLLM Evaluation & TestingAI Guardrails & Content SafetySpeech Recognition & TranscriptionExperiment Tracking & ML OpsDocument AI & OCR

About

How we rank & score

Layer 3 · self-hosting reality check

What it actually takes to self-host LocalAI

The docs say 4 GB. In practice you want 8 GB RAM for CPU inference; 8 GB VRAM for anything comfortable. Here is the honest version — real requirements, real monthly cost, what you will be maintaining, and the one thing that catches people out.

Usually reached from OpenAI API (ChatGPT) alternatives, where LocalAI is one of the picks.

RAM — documented minimum4 GB
RAM — what it really needs8 GB RAM for CPU inference; 8 GB VRAM for anything comfortable
CPU8 vCPU if running on CPU
Disk4–40 GB per model
Monthly cost$0 on hardware you own; a rented GPU is $150–400/mo
Setup time1 hour
How you install itcontainer; it reimplements the OpenAI API surface across chat, embeddings, images and audio
Ongoing maintenanceModerate — broad backend support means a broad configuration surface.
Where it stops scalingSingle-user and small-team use. Breadth, not concurrency, is the design.

The thing that catches people out

It replaces the whole OpenAI API surface rather than just chat, which is the appeal and the trap: quality varies enormously depending on which local model backs each endpoint. An excellent chat model paired with a weak embedding model produces a RAG pipeline that fails in ways that look like retrieval bugs. Evaluate each endpoint you actually use, separately.

When not to self-host LocalAI

You need throughput under concurrency. vLLM is built for serving; LocalAI is built for breadth.

Every guide here carries this section. A site that only ever tells you to self-host is selling something — the useful answer is sometimes no.

Other Layer 3 self-hosting guides

Common questions

How much RAM does LocalAI actually need?
8 GB RAM for CPU inference; 8 GB VRAM for anything comfortable in practice. The documented minimum is 4 GB, which is the figure at which the process starts rather than the figure at which it works under real use. 8 vCPU if running on CPU alongside it.
What does self-hosting LocalAI cost per month?
$0 on hardware you own; a rented GPU is $150–400/mo This is commodity VPS pricing and excludes your time, which is the larger cost for most people — budget for moderate — broad backend support means a broad configuration surface.
How long does it take to set up LocalAI?
1 hour, via container; it reimplements the OpenAI API surface across chat, embeddings, images and audio.
When should I NOT self-host LocalAI?
You need throughput under concurrency. vLLM is built for serving; LocalAI is built for breadth.
What is the most common mistake when self-hosting LocalAI?
It replaces the whole OpenAI API surface rather than just chat, which is the appeal and the trap: quality varies enormously depending on which local model backs each endpoint. An excellent chat model paired with a weak embedding model produces a RAG pipeline that fails in ways that look like retrieval bugs. Evaluate each endpoint you actually use, separately.
The Macrostack brief

New swaps, worth your inbox.

A short, occasional email when we add a high-intent alternative or ship a new head-to-head. No spam, no selling your address — unsubscribe in one click.