n8500x

Next-token-logit multiple-choice classifier (Jev in 25 lines) as offline API+CLI. CPU, modelless.

Static Scala -> Python converter, verified by scalac-vs-python differential runs

Customized Jenkins LTS with baked-in CLI tools, ~150 plugins, and prefetched Maven artifacts

Claude Code on-prem: OpenAI/Anthropic gateways (API key, no-auth or OIDC) + Bedrock.

Patched Jenkins inbound (JNLP) agent on Alpine, hardened for OpenShift restricted-v2

Open-source Kafka Connect 4.3.1: Redis Streams, JDBC, OpenSearch, MongoDB, ActiveMQ

RHEL 9 GPU node kit: NVIDIA driver local repo, Container Toolkit, dkms, k8s device plugin, runbook

CP Kafka Connect Redis Connector

llama.cpp llama-server from source: CPU-only, portable variants, modelless, OpenAI-compatible

HAProxy (lts-alpine) with Lua + LuaSec (ssl.https) for HTTPS from Lua scripts

HAProxy (lts-alpine) with Lua + LuaSec (ssl.https) for HTTPS from Lua scripts

Chat any backend: Kafka, k8s/OCP, Postgres, MSSQL, Oracle, Redis, Mongo, AWS/Azure, Cloudera

Air-gapped Kafka 3.7 + ZooKeeper installer: vendored charts, SASL_SSL/SCRAM, 1-command Argo CD

AI root-cause analysis for SQL Server, on any model endpoint (LiteLLM/vLLM/OpenAI/Anthropic).

Grafana + ClickHouse datasource (baked) for the SQL Server RCA Assistant trends (optional).

DMV collector for the SQL Server RCA Assistant monitoring stack (optional companion image).

DEPRECATED - use n8500x/mcp-chat-all (single image, all backends)

Latest official opencode for autonomous CI/headless - JDK17/21+Gradle8.8+Py3.14+Node24

Standalone Prometheus JMX Exporter - scrapes a JVM over JMX, serves metrics on :8080

Self-contained Terraform + AWS CLI image that provisions and drives an AWS LLM-inference MLOps stack

Self-contained Terraform + Azure CLI runner that provisions a private Azure OpenAI + LiteLLM stack

DevOps Swiss Army Knife (`devops-sak`)

OpenMetadata deployment

Air-gapped Argo CD Helm chart image (chart 9.5.11 / Argo CD v3.3.9); helm entrypoint

LiteLLM Proxy — Air-Gapped Production Image

Air-gapped LLM inference benchmarking tool (llama-benchy) on python:3.12-slim

Qwen Code CLI agent container, with an Ollama backend via docker-compose

Offline/air-gapped aider AI coding-assistant CLI (Python 3.12) with a full dev toolbelt

Kafka MCP Server