{
 "slug": "localai",
 "name": "LocalAI",
 "tagline": "OpenAI-compatible API server for running LLMs, speech and image models on CPU",
 "category": "ai-llm",
 "use_cases": [
  "openai-compatible-api",
  "llm-inference",
  "private-llm"
 ],
 "aliases": [],
 "status": "live",
 "upstream": {
  "vendor": "Ettore Di Giacinto",
  "url": "https://localai.io",
  "license": "MIT",
  "license_url": "https://github.com/mudler/LocalAI/blob/master/LICENSE"
 },
 "image": {
  "os": "Ubuntu 24.04 LTS",
  "cis_level": 1,
  "ports": [
   80,
   443
  ],
  "artifacts": [
   "/etc/lynxroute/sbom.json",
   "/etc/lynxroute/cis-report.html",
   "/usr/share/doc/lynxroute/CIS_TAILORED_PROFILE.md"
  ]
 },
 "clouds": {
  "aws": {
   "status": "live",
   "version": "4.12.1",
   "last_build": "2026-10-25",
   "entity_id": "prod-lfouziothu6du",
   "ami_alias": "/aws/service/marketplace/prod-lfouziothu6du/latest",
   "listing_url": "https://aws.amazon.com/marketplace/pp/prodview-5usomzm5h3qj6",
   "recommended_instance_type": "t3.large",
   "price_usd_per_hour": 0.05,
   "free_trial_days": 5,
   "unit": "instance-hour"
  },
  "azure": {
   "status": "live",
   "version": "4.10.0",
   "last_build": "2026-09-20",
   "offer_id": "localai",
   "plan_id": "payg-gen2",
   "image_urn": "lynxroute:localai:payg-gen2:latest",
   "listing_url": "https://azuremarketplace.microsoft.com/en-us/marketplace/apps/lynxroute.localai?tab=Overview",
   "recommended_vm_size": "Standard_D2s_v3",
   "price_usd_per_vcpu_hour": 0.028,
   "unit": "vcpu-hour"
  },
  "alicloud": {
   "status": "not_offered"
  }
 },
 "urls": {
  "page": "https://lynxroute.com/products/localai/",
  "json": "https://lynxroute.com/products/localai/product.json"
 },
 "release_notes": [
  {
   "version": "4.12.1",
   "title": "LocalAI 4.12.1",
   "notes": "* LocalAI updated 4.12.0 -> 4.12.1, a security-only patch release. MIT unchanged\n* Security: hardened handling of user-supplied URLs (vision, image and audio endpoints): every redirect hop is validated, carrier-grade NAT and 0.0.0.0/8 ranges are treated as non-public, and the connection goes to the validated address\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base, picking up every operating-system security update published since the previous build\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-10-25"
   }
  },
  {
   "version": "4.12.0",
   "title": "LocalAI 4.12.0",
   "notes": "* LocalAI updated 4.10.0 -> 4.12.0. MIT unchanged\n* Redesigned web UI, streaming voice, PII controls, model failover and MCP OAuth2 support\n* Security: validates the CSRF-exemption header and rejects oversized GGUF model metadata\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base, picking up every operating-system security update published since the previous build\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-10-23"
   }
  },
  {
   "version": "4.10.0",
   "title": "LocalAI 4.10.0",
   "notes": "* LocalAI 4.10.0 on Ubuntu 24.04 LTS - minor update (from 4.9.0). MIT unchanged\n* Security: four CVEs patched in bundled dependencies (ip-address, containerd, react-router)\n* Credentials file - one credentials.yaml authenticates private OCI registries, galleries, HTTP downloads and github: URIs with basic, bearer or custom-header auth, read at use time and masked in logs; a credential is never carried across a redirect to another host\n* Per-model environment variables - an env: map in a model config injects variables such as CUDA_VISIBLE_DEVICES into that model's backend process, with no wrapper script\n* Fleet operations dashboard replaces the flat node list: cluster health bands, VRAM/RAM/CPU/disk capacity gauges, a running-models view with replica placement, and bulk drain, resume and remove\n* New \"local-ai benchmark\" command reports end-to-end latency and completion tokens per second for configured text models, as a table or JSON\n* template.system_messages_after_first fixes agent frameworks that append system instructions mid-conversation, which previously returned HTTP 500 on Qwen3.x models\n* /v1/models/capabilities now reports each model's effective context window, so clients can budget tokens without reading the model config\n* 140 new gallery entries; existing model configurations and API routes are unchanged\n* Certbot pre-installed - enable a trusted HTTPS certificate with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-09-20",
    "azure": "2026-09-20"
   }
  },
  {
   "version": "4.9.0",
   "title": "LocalAI v4.9.0",
   "notes": "* Log rotation no longer restarts the API server. The previous logrotate config signalled local-ai with SIGHUP, which it treats as a shutdown request; systemd restarted it, but every rotation dropped all loaded models and in-flight requests and forced a cold model reload. Rotation now uses `copytruncate` and signals nothing. LocalAI 4.9.0 itself is unchanged.\n* Rebuilt on the current CIS Level 1 hardened Ubuntu 24.04 LTS base.\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-08-24",
    "azure": "2026-08-26"
   }
  },
  {
   "version": "4.9.0",
   "title": "LocalAI 4.9.0",
   "notes": "* LocalAI 4.9.0 on Ubuntu 24.04 LTS - minor update (from 4.8.2). MIT unchanged\n* Security: authentication is deny-by-default - every HTTP route requires credentials unless it appears in an explicit public registry, closing a bypass in which unprefixed aliases such as /moderations, /models, /backends and /mcp/chat/completions fell outside the old protected-prefix list\n* Chat context compression, opt-in per model: older complete turns are compressed before inference, preserving system prompts, the newest messages and whole tool-call units\n* One page per resource - /app/models owns Explore and Installed, /app/backends owns Catalog and Installed; existing /app/manage bookmarks still work\n* Reversible request-scoped PII pseudonyms, global HTTP admission control with live backend traces, and parallel Hugging Face downloads\n* Certbot pre-installed - enable a trusted HTTPS certificate with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-08-24",
    "azure": "2026-08-26"
   }
  },
  {
   "version": "4.8.2",
   "title": "LocalAI 4.8.2 on Ubuntu 24.04 LTS",
   "notes": "* LocalAI 4.8.2 on Ubuntu 24.04 LTS - minor update (from 4.7.1). MIT unchanged\n* The model gallery now falls back to mirrors and a cached index when the primary source is unavailable, so model browsing keeps working through an upstream outage\n* Certbot pre-installed - enable a trusted HTTPS certificate with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-git",
   "live": {
    "aws": "2026-08-18",
    "azure": "2026-08-19"
   }
  },
  {
   "version": "4.7.1",
   "title": "LocalAI 4.7.1 on Ubuntu 24.04 LTS",
   "notes": "* LocalAI 4.7.1 on Ubuntu 24.04 LTS - minor update (from 4.6.2): voice-cloning profile endpoints, auto-resolve context size, GPU device selection, and inbound thinking/reasoning aliases - all additive, existing options unchanged. MIT unchanged\n* Certbot pre-installed - enable a trusted HTTPS certificate with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-git",
   "live": {
    "aws": "2026-07-20",
    "azure": "2026-07-20"
   }
  },
  {
   "version": "4.6.2",
   "title": "LocalAI 4.6.2 on Ubuntu 24.04 LTS",
   "notes": "* LocalAI 4.6.2 on Ubuntu 24.04 LTS - minor update (from 4.5.0): realtime pipeline warm-up, a model capabilities endpoint, added Prometheus metrics, chat-UI improvements, and process-lifecycle fixes. MIT unchanged\n* Certbot pre-installed - enable a trusted HTTPS certificate with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-07-11",
    "azure": "2026-07-11"
   }
  },
  {
   "version": "4.5.0",
   "title": "LocalAI 4.5.0 - June 2026",
   "notes": "* LocalAI 4.5.0 on Ubuntu 24.04 LTS - feature release (minor from 4.4.3)\n* New model backends, redesigned Web UI, distributed-mode robustness\n* Tuned performance defaults: prefix caching enabled, VRAM-scaled parallelism\n* Certbot pre-installed - enable a trusted HTTPS certificate: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-06-25"
   }
  },
  {
   "version": "4.4.3",
   "title": "LocalAI 4.4.3",
   "notes": "* Updated LocalAI to 4.4.3\n* Certbot pre-installed - enable HTTPS with one command: sudo certbot --nginx -d yourdomain.com\n* Rebuilt on the latest CIS Level 1 hardened Ubuntu 24.04 LTS base\n",
   "source": "backfill-git",
   "live": {
    "aws": "2026-06-16",
    "azure": "2026-06-20"
   }
  },
  {
   "version": "4.2.6",
   "title": "LocalAI 4.2.6 - May 2026",
   "notes": "* LocalAI 4.2.6 on Ubuntu 24.04 LTS\n* llama-cpp bumped to MTP-merge SHA with automatic MTP defaults\n* Dependency updates: antirez/ds4, ikawrakow/ik_llama.cpp, leejet/stable-diffusion.cpp\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-05-23"
   }
  },
  {
   "version": "4.2.5",
   "title": "LocalAI 4.2.5 - Initial release (May 2026)",
   "notes": "* LocalAI 4.2.5 single-binary on Ubuntu 24.04 LTS\n* CIS Level 1 hardening applied (ansible-lockdown/UBUNTU24-CIS)\n* CVE-scanned before every release\n* Per-instance LOCALAI_API_KEY (24-char random) generated at first boot - enforced by LocalAI for all /v1/, /models, /embeddings, /audio, /images calls\n* LocalAI bound to 127.0.0.1:8080 and reachable only through host Nginx with TLS\n* Same API key authenticates OpenAI-SDK REST clients (Authorization: Bearer) and the Web UI (\"Login with API Token\")\n* UFW firewall pre-configured (TCP 22, 80, 443 only)\n* fail2ban, auditd, AppArmor pre-configured\n* SBOM (CycloneDX 1.6) at /etc/lynxroute/sbom.json\n* CIS Conformance Report (OpenSCAP) at /etc/lynxroute/cis-report.html\n* IMDSv2 enforced\n",
   "source": "backfill-aws-listing",
   "live": {
    "aws": "2026-05-19"
   }
  }
 ],
 "release_history": [
  {
   "version": "4.12.1",
   "aws": "2026-10-25"
  },
  {
   "version": "4.12.0",
   "aws": "2026-10-23"
  },
  {
   "version": "4.10.0",
   "aws": "2026-09-20",
   "azure": "2026-09-20"
  },
  {
   "version": "4.9.0",
   "aws": "2026-08-24",
   "azure": "2026-08-26"
  },
  {
   "version": "4.8.2",
   "aws": "2026-08-18",
   "azure": "2026-08-19"
  },
  {
   "version": "4.7.1",
   "aws": "2026-07-20",
   "azure": "2026-07-20"
  },
  {
   "version": "4.6.2",
   "aws": "2026-07-11",
   "azure": "2026-07-11"
  },
  {
   "version": "4.5.0",
   "aws": "2026-06-25"
  },
  {
   "version": "4.4.3",
   "aws": "2026-06-16",
   "azure": "2026-06-20"
  },
  {
   "version": "4.3.6",
   "aws": "2026-06-05",
   "azure": "2026-06-07"
  },
  {
   "version": "4.2.6",
   "aws": "2026-05-23"
  },
  {
   "version": "4.2.5",
   "aws": "2026-05-19"
  }
 ],
 "deploy": {
  "aws_cli": "# subscribe once on the AWS Marketplace listing, then:\naws ec2 run-instances \\\n  --image-id resolve:ssm:/aws/service/marketplace/prod-lfouziothu6du/latest \\\n  --instance-type t3.large --key-name my-key --security-group-ids sg-xxxxxxxx",
  "azure_cli": "# accept the plan terms once, then:\naz vm image terms accept --publisher lynxroute --offer localai --plan payg-gen2\naz vm create -g my-rg -n localai --size Standard_D2s_v3 \\\n  --image lynxroute:localai:payg-gen2:latest --admin-username azureuser --generate-ssh-keys"
 },
 "description": {
  "what_is": "LocalAI is an open-source, OpenAI-compatible inference server delivered as a single statically linked Go binary. It exposes the same REST surface as OpenAI - /v1/chat/completions, /v1/embeddings, /v1/audio/transcriptions, /v1/audio/speech, /v1/images/generations, /v1/models, plus a built-in /models gallery API - so any client written for OpenAI (openai-python, openai-node, LangChain OpenAI provider, LlamaIndex, IDE plugins) talks to it unchanged with one base URL swap. Inference runs entirely on CPU using llama.cpp and Whisper.cpp backends; GGUF model files load from local disk under /var/lib/localai/models and persist across restarts - no external database, no Redis, no cloud egress to OpenAI or Anthropic. Bring your own GGUF model (LLaMA family, Mistral, Qwen, Phi, Gemma) or install one in two clicks from the bundled gallery (mudler/LocalAI gallery index). MIT license, no vendor lock-in.",
  "image_adds": "Security hardening:\n* Per-instance LOCALAI_API_KEY (24-char random) generated at first boot, enforced by LocalAI for every /v1/, /models, /embeddings, /audio, /images call - never baked into the image\n* The same API key authenticates OpenAI-SDK REST clients (Authorization: Bearer) and the Web UI (\"Login with API Token\") - one credential, one rotation surface\n* LocalAI bound to 127.0.0.1:8080 - the API process never listens on a public interface\n* Host Nginx fronts port 443 with a self-signed certificate; HTTP redirects to HTTPS; security headers (X-Content-Type-Options, X-Frame-Options, Referrer-Policy) applied\n* UFW firewall pre-configured - only TCP 22, 80, 443 are exposed\n* fail2ban, AppArmor\n* CVE scan - every image is scanned for vulnerabilities before release\n\nOS hardening (CIS Level 1):\n* CIS Ubuntu 24.04 LTS Level 1 benchmark applied via ansible-lockdown\n* auditd, SSH hardening, kernel hardening, IMDSv2 enforced\n\nCompliance artifacts:\n* SBOM - CycloneDX 1.6 at /etc/lynxroute/sbom.json\n* CIS Conformance Report at /etc/lynxroute/cis-report.html\n* CIS Tailored Profile at /usr/share/doc/lynxroute/CIS_TAILORED_PROFILE.md"
 },
 "generated_at": "2026-10-31T09:36:34Z"
}