> ## Documentation Index > Fetch the complete documentation index at: https://docs.neuraltrust.ai/llms.txt > Use this file to discover all available pages before exploring further. # Image catalog > Complete inventory of container images the NeuralTrust Platform Helm chart deploys, grouped by subchart and by deployment model. Use this when mirroring images, planning air-gapped installs, or sizing image-pull bandwidth. All NeuralTrust images live in the private Google Artifact Registry\ `europe-west1-docker.pkg.dev/neuraltrust-app-prod/nt-docker/`. To mirror or rebrand the registry, set `global.imageRegistry` once and the chart will rewrite all subchart image paths automatically (the helper strips the default GCP prefix before prepending yours). All private images are pulled with `gcr-secret` (`docker-registry` type). The Kubernetes Secret name is configurable per subchart via `imagePullSecrets`. Versions below reflect the current `values.yaml` defaults in the chart. They change with every chart release — always cross-check against the `values.yaml` of the chart version you're installing. ## Per-subchart inventory ### `clickhouse` | Image | Default tag | Purpose | Default replicas | Storage | | --------------------- | ----------- | ------------------------------------------------------- | ---------------- | ---------- | | `…/clickhouse-server` | `26.3` | ClickHouse DB (StatefulSet) | 1 | 50 GiB PVC | | `…/clickhouse-server` | `26.3` | Backup CronJob (when `clickhouse.backup.enabled: true`) | CronJob | — | ### `kafka` | Image | Default tag | Purpose | Default replicas | Storage | | --------- | ----------- | -------------------------- | ---------------- | ---------- | | `…/kafka` | `4.3.0` | Kafka broker (StatefulSet) | 1 | 10 GiB PVC | ### `neuraltrust-data-plane` | Image | Default tag | Purpose | Default replicas | | --------------------- | ----------- | --------------------------------------------------------------------------------------------------------------- | ---------------- | | `…/data-plane-api` | `v1.24.11` | Data Plane API | 2 | | `…/clickhouse-server` | `26.3` | Init container `clickhouse-migrations` on API | — | | `…/workers` | `v1.6.12` | Data Plane worker | 1 | | `…/kafka-connect` | `v0.3.1` | Kafka Connect Deployment | 1 | | `…/kafka-connect` | `v0.3.1` | Init container `fix-connect-topics` on Kafka Connect | — | | `…/data-plane-api` | `v1.24.11` | Ephemeral Kubernetes Job pods spawned by API for evals (`k8sJobs.enabled: true`, up to `maxConcurrentJobs: 10`) | — | ### `neuraltrust-control-plane` *(only when `controlPlane.enabled: true`)* | Image | Default tag | Purpose | Default replicas | | --------------------- | ----------- | ----------------------------------------------------------------------- | ---------------- | | `…/control-plane-api` | `v1.18.3` | Control Plane API | 2 | | `…/app` | `v1.65.9` | Control Plane UI (Next.js) | 2 | | `…/app` | `v1.65.9` | Init container `init-db` (Prisma migrations) on UI | — | | `…/scheduler` | `v1.9.7` | Scheduler | 1 | | `…/postgres` | `17-alpine` | Init container `wait-for-postgresql` on API + scheduler | — | | `…/postgres` | `17-alpine` | In-cluster PostgreSQL *(when `infrastructure.postgresql.deploy: true`)* | 1 | ### `trustgate` | Image | Default tag | Purpose | Default replicas | | ---------------------- | ----------- | -------------------------------------------------------- | ---------------- | | `…/trustgate-ee` | `v1.27.5` | TrustGate admin (`trustgate-control-plane`) | 2 | | `…/trustgate-ee` | `v1.27.5` | TrustGate gateway (`trustgate-data-plane`) | 2 | | `…/trustgate-ee` | `v1.27.5` | TrustGate actions (`trustgate-actions`) | 2 | | `…/postgres` | `17-alpine` | Init Job `trustgate-postgresql-init` (creates DB + user) | Job | | `…/redis-stack-server` | `7.2.0-v20` | Redis cache | 1 | ### `neuraltrust-firewall` *(default: enabled)* | Image | Default tag | Purpose | Default replicas | | ---------------- | ----------- | ----------------------------------------------------------- | ---------------- | | `…/firewall-cpu` | `v2.9.6` | Gateway (CPU router) | 2 | | `…/firewall-cpu` | `v2.9.6` | Worker — `toxicity` | 1 | | `…/firewall-cpu` | `v2.9.6` | Worker — `toolguard` | 1 | | `…/firewall-cpu` | `v2.9.6` | Worker — `prompt-jailbreak` | 1 | | `…/firewall-cpu` | `v2.9.6` | Worker — `prompt-moderation` | 1 | | `…/firewall-cpu` | `v2.9.6` | Worker — `response-jailbreak` | 1 | | `…/firewall-gpu` | `v2.9.6` | GPU-accelerated workers *(override `workerDefaults.image`)* | configurable | ### Optional add-on subcharts *(off by default)* | Subchart | Image(s) | Default tag | Toggle | | -------------------------------- | -------------------------------------- | ------------------ | ---------------------------------------------------- | | `neuraltrust-siem-connectors` | `siem-connectors` | `v0.2.2` | `neuraltrust-siem-connectors.siemConnectors.enabled` | | `neuraltrust-watchdog` | `neuraltrust-watchdog` | Chart `appVersion` | `neuraltrust-watchdog.enabled` | | Umbrella OpenTelemetry Collector | `otel/opentelemetry-collector-contrib` | `0.110.0` | `global.observability.enabled` | ## What's deployed in each model The two main deployment models are documented in [Deployment models](./deployment-models). The image footprint for each: Chart default. Control Plane runs on NeuralTrust SaaS — no CP images in your cluster. | Subchart | Deployed? | Notes | | ----------------------------------------------- | ------------------- | ----------------------------------------------------- | | `neuraltrust-data-plane` | ✅ | API ×2 + worker ×1 + Kafka Connect ×1 | | `trustgate` | ✅ *(typical)* | admin ×2, gateway ×2, actions ×2, Redis ×1 | | `neuraltrust-firewall` | ✅ | gateway ×2 + 5 workers (CPU by default, GPU optional) | | `clickhouse` | ✅ *(or external)* | StatefulSet ×1 | | `kafka` | ✅ *(or external)* | StatefulSet ×1 | | `neuraltrust-control-plane` PostgreSQL | ✅ *(for TrustGate)* | StatefulSet ×1 | | `neuraltrust-control-plane` CP API/UI/Scheduler | ❌ | runs on NeuralTrust SaaS | **Distinct in-cluster images (defaults):** * `data-plane-api`, `workers`, `kafka-connect` * `trustgate-ee`, `redis-stack-server`, `postgres` (init Job) * `firewall-cpu` (gateway + 5 worker pods) * `clickhouse-server`, `kafka`, `postgres` Total: \~10 distinct image repositories, \~16 running pods on chart defaults. Adds Control Plane API, UI, and Scheduler on top of the hybrid footprint. | Subchart | Deployed? | Notes | | ------------------------------------------------- | ----------------- | -------------------------------- | | `neuraltrust-data-plane` | ✅ | same as hybrid | | `neuraltrust-control-plane` (CP API/UI/Scheduler) | ✅ | API ×2, UI ×2, Scheduler ×1 | | `neuraltrust-control-plane` PostgreSQL | ✅ *(or external)* | shared with TrustGate by default | | `trustgate` | ✅ | same as hybrid | | `neuraltrust-firewall` | ✅ | same as hybrid | | `clickhouse` | ✅ *(or external)* | same as hybrid | | `kafka` | ✅ *(or external)* | same as hybrid | **Additional images vs. hybrid:** * `control-plane-api`, `app`, `scheduler` Total: \~13 distinct image repositories, \~21 running pods on chart defaults. ## Sizing & resource defaults Per-component requests / limits (from chart `values.yaml`). Multiply by the replica count to get the per-component total contribution to cluster requests. | Component | Replicas | CPU req / limit | Memory req / limit | GPU | | -------------------------------------------------------------------------------------------------- | -------- | --------------------------------- | ------------------ | ----- | | ClickHouse | 1 | 2 / 4 | 4 Gi / 8 Gi | — | | Kafka | 1 | 500m / 1 | 1 Gi / 2 Gi | — | | PostgreSQL | 1 | 1 / 2 | 2 Gi / 4 Gi | — | | Data Plane API | 2 | 2 / 4 | 4 Gi / 6 Gi | — | | Data Plane worker | 1 | 1 / 2 | 4 Gi / 8 Gi | — | | Kafka Connect | 1 | 500m / 1 | 2 Gi / 4 Gi | — | | CP scheduler | 1 | 100m / 200m | 256 Mi / 512 Mi | — | | CP API | 2 | 1 / 2 | 1 Gi / 2 Gi | — | | CP UI (app) | 2 | 250m / 500m | 512 Mi / 1 Gi | — | | TrustGate admin | 2 | 500m / 1 | 2 Gi / 4 Gi | — | | TrustGate gateway | 2 | 2 / 4 | 4 Gi / 8 Gi | — | | TrustGate actions | 2 | 500m / 1 | 2 Gi / 4 Gi | — | | Redis | 1 | 500m / 1 | 1 Gi / 2 Gi | — | | Firewall gateway | 2 | 250m / 500m | 256 Mi / 512 Mi | — | | Firewall worker (×5: toxicity, toolguard, prompt-jailbreak, prompt-moderation, response-jailbreak) | 5 | 1 / 1.5 | 4 Gi / 6 Gi | — | | Firewall GPU worker *(replaces CPU worker)* | per pool | same as CPU + `nvidia.com/gpu: 1` | same | 1 GPU | | K8s eval Job *(ephemeral, spawned by Data Plane API)* | up to 10 | 100m / 500m | 256 Mi / 1 Gi | — | ### Realistic cluster totals Firewall is always deployed — the choice is whether its 5 default workers run on the **CPU pool** or move to a **separate GPU pool**. | Model | CPU pool requests | PVC | Min CPU nodes *(8 vCPU / 32 GiB)* | GPU pool | | ------------------------------------- | ------------------------ | -------- | --------------------------------- | ------------ | | **Hybrid + CPU Firewall** *(default)* | \~20.5 vCPU / \~58.5 GiB | \~80 GiB | **≥ 4** | — | | **Hybrid + GPU Firewall** | \~15.5 vCPU / \~38.5 GiB | \~80 GiB | **≥ 3** | ≥ 5 GPU pods | | **Self-hosted + CPU Firewall** | \~23.1 vCPU / \~61.8 GiB | \~80 GiB | **≥ 5** | — | | **Self-hosted + GPU Firewall** | \~18.1 vCPU / \~41.8 GiB | \~80 GiB | **≥ 4** | ≥ 5 GPU pods | CPU Firewall workers contribute 5 vCPU + 20 GiB to the main pool. Switching to GPU workers moves that load to dedicated GPU nodes (one GPU per worker without CUDA MPS). For per-cloud node SKUs (e.g. `e2-standard-8`, `m5.2xlarge`, `Standard_D8s_v5`, `g4dn.xlarge` for GPU) and HA layout, see the cloud-specific overview pages and [Deployment models › Sizing baseline](./deployment-models#sizing-baseline). ## Mirroring images for air-gapped installs To pre-pull every image the chart would deploy with your overrides: ```bash theme={null} helm template neuraltrust-platform \ oci://europe-west1-docker.pkg.dev/neuraltrust-app-prod/helm-charts/neuraltrust-platform \ --version \ -f my-values.yaml \ | yq '.. | select(has("image")) | .image' \ | sort -u ``` Then mirror each image to your internal registry and set: ```yaml theme={null} global: imageRegistry: "registry.internal.example.com/neuraltrust" ``` The chart strips the default `europe-west1-docker.pkg.dev/neuraltrust-app-prod/nt-docker` prefix from subchart image paths and replaces it with your registry. Per-subchart `image.repository` overrides still win if set explicitly. ## Pull policy | Component | Default pull policy | | ------------------------------------------ | ---------------------------- | | Data Plane API, worker, Kafka Connect | `Always` | | Control Plane API, UI, Scheduler, Postgres | `Always` | | TrustGate admin, gateway, actions | `Always` | | TrustGate Redis | `Always` *(parent override)* | | Firewall gateway + workers | `IfNotPresent` | | ClickHouse, Kafka | `IfNotPresent` | `Always` ensures pinned tags still revalidate on rollouts. Override per component if you mirror to an internal registry that doesn't change tags. ## Related guides * [Deployment models](./deployment-models) — hybrid vs self-hosted in depth * [Configuration scenarios](./configuration) — values files for each topology * [Secrets management](./secrets) — secrets each image needs * [Firewall deployment](./firewall) — CPU vs GPU, per-worker overrides