GitHub Repository

Docker image to run an Ollama local LLM server. Secure by default, all API requests require a Bearer token (auto-generated on first start). OpenAI-compatible API. Supports first-start model pre-pull, NVIDIA GPU (CUDA) acceleration, and persistent model storage. Multi-arch: amd64, arm64.

7 starsShell

Secure-by-default Ollama Docker image with built-in auth, only 70MB

Name: Secure-by-default Ollama Docker image with built-in auth, only 70MB
Availability: InStock
Author: hwdsl2

by hwdsl2·May 1, 2026·1 point·0 comments

Visit Project View on HN

AI Analysis

●●SolidSolve My ProblemShip It

Auto-generated tokens block the 175k exposed Ollama servers found online.

Strengths

•Enforces Bearer token auth via Caddy proxy without manual config.
•Tiny 70MB image size compared to standard Ollama Docker builds.
•Auto-pulls models on first start via environment variable configuration.

Weaknesses

•No built-in HTTPS termination, requires external reverse proxy for TLS.
•Still relies on Ollama upstream for actual model inference performance.

Similar Projects

Infrastructure●●Solid

Secure-by-default Ollama Docker image with built-in auth, only ~70MB

Auto-generates API keys to block the 175k exposed Ollama instances.

Solve My ProblemCozy

hwdsl2

201mo ago

Security●●Solid

NixOS flake for hardened OpenClaw deployment

Two lines in your flake flip OpenClaw from alarmingly exposed to locked-down: gateway auth, localhost binding, Caddy auto-TLS, strict systemd directives, tool allowlists, and fail2ban are all wired in. It's a pragmatic, opinionated safety wrapper that saves you from the default footguns — just expect it to be useful only if you already live in the NixOS/OpenClaw world.

Niche GemSolve My Problem

scout_agent

104mo ago

Developer Tools●●Solid

Nuvix – An Open Source Back End Where Every Table Is Secure by Default

Auto-generated Row-Level Security plus a three-schema model (document / managed / unmanaged) is the concrete feature that makes this more than yet another BaaS rewrite — you can prototype like NoSQL then flip to managed tables with permission tables and RLS without hand-writing policies. It bundles auth, permission-aware file storage, messaging and embeddings storage which is useful for ML workflows, but it's entering a crowded field (Supabase/Hasura/Appwrite) and would benefit from clearer migration stories, performance numbers, and a hosted roadmap.

Ship ItNiche Gem

ravikantsaini

104mo ago

AI/ML●●Solid