docs: add VoltageGPU custom endpoint guide

components/icons/providers/index.ts

@@ -47,6 +47,7 @@ export { default as TogetherAIIcon } from './togetherai'
 export { default as TraefikIcon } from './traefik'
 export { default as TrueFoundryIcon } from './truefoundry'
 export { default as VLLMIcon } from './vllm'
+export { default as VoltageGPUIcon } from './voltagegpu'
 export { default as VultrIcon } from './vultr'
 export { default as WolframIcon } from './wolfram'
 export { default as XAIIcon } from './xai'

components/icons/providers/voltagegpu.tsx

@@ -0,0 +1,29 @@
+import React from 'react'
+
+// VoltageGPU mark (ring and bolt), from the logo component of voltagegpu.com
+export default function VoltageGPUIcon({ className, ...props }: React.SVGProps<SVGSVGElement>) {
+  return (
+    <svg
+      xmlns="http://www.w3.org/2000/svg"
+      viewBox="0 0 32 32"
+      fill="none"
+      className={className}
+      aria-hidden="true"
+      {...props}
+    >
+      <path
+        d="M16 2C8.268 2 2 8.268 2 16C2 23.732 8.268 30 16 30C23.732 30 30 23.732 30 16C30 8.268 23.732 2 16 2Z"
+        stroke="currentColor"
+        strokeWidth="2"
+        fill="none"
+      />
+      <path
+        d="M16 8L10 18H16L16 24L22 14H16L16 8Z"
+        fill="currentColor"
+        stroke="currentColor"
+        strokeWidth="0.5"
+        strokeLinejoin="round"
+      />
+    </svg>
+  )
+}

content/docs/configuration/librechat_yaml/ai_endpoints/meta.json

@@ -28,6 +28,7 @@
     "togetherai",
     "truefoundry",
     "vllm",
+    "voltagegpu",
     "vultrcloudinference",
     "xai"
   ]

content/docs/configuration/librechat_yaml/ai_endpoints/voltagegpu.mdx

@@ -0,0 +1,44 @@
+---
+title: VoltageGPU
+icon: VoltageGPU
+description: Configure VoltageGPU as a custom endpoint in LibreChat.
+---
+
+VoltageGPU provides an OpenAI-compatible endpoint that you can add to LibreChat through `endpoints.custom`. Its models run in a trusted execution environment (Intel TDX guests with confidential-computing GPUs), and every model ID ends in `-TEE`. Requests pass through VoltageGPU's relay, which sees them in clear (TLS ends at VoltageGPU's host). To keep plaintext off VoltageGPU's servers, run your own inference server in a [Confidential VM](https://voltagegpu.com/confidential-vm) and terminate TLS inside its trust domain.
+
+## Get an API key
+
+Create an API key from the [VoltageGPU API keys page](https://voltagegpu.com/api-keys). Add it to your `.env` file:
+
+```bash filename=".env"
+VOLTAGEGPU_API_KEY=your-api-key
+```
+
+## Configuration
+
+Add the endpoint to the `custom` list under `endpoints` in your `librechat.yaml`. If your file already has an `endpoints.custom` list, paste only the `- name:` entry into it, keeping the same indentation:
+
+```yaml filename="librechat.yaml"
+endpoints:
+  custom:
+    - name: "VoltageGPU"
+      apiKey: "${VOLTAGEGPU_API_KEY}"
+      baseURL: "https://api.voltagegpu.com/v1"
+      models:
+        default:
+          - "Qwen/Qwen3.8-27B-TEE"
+          - "deepseek-ai/DeepSeek-V3.2-TEE"
+        fetch: true
+      titleConvo: true
+      titleModel: "Qwen/Qwen3.8-27B-TEE"
+      modelDisplayLabel: "VoltageGPU"
+```
+
+To let each user supply their own key through the LibreChat UI instead of reading one from `.env`, set `apiKey: "user_provided"`.
+
+## Notes
+
+- With `fetch: true`, LibreChat loads the model list from VoltageGPU's OpenAI-compatible `/v1/models` endpoint. The `default` array is only the initial selection.
+- Use model IDs exactly as they appear in the [model list](https://api.voltagegpu.com/v1/models). The API only accepts IDs ending in `-TEE`.
+- Each entry in `/v1/models` carries a live `status.state` (`available`, `busy`, `saturated` or `unavailable`) and its price per million tokens. If a selected model is unavailable, choose another chat-capable model from the list.
+- `GET /v1/attestation?model=<id>&nonce=<64 hex characters>` returns the Intel TDX quotes and NVIDIA GPU evidence of that model's running instances, bound to your nonce, for you to verify. See the [VoltageGPU documentation](https://docs.voltagegpu.com) for details.

lib/icons.tsx

@@ -49,6 +49,7 @@ import {
   TraefikIcon,
   TrueFoundryIcon,
   VLLMIcon,
+  VoltageGPUIcon,
   VultrIcon,
   WolframIcon,
   XAIIcon,
@@ -285,6 +286,7 @@ const icons: Record<string, ReactElement> = {
   Traefik: <TraefikIcon />,
   TrueFoundry: <TrueFoundryIcon />,
   VLLM: <VLLMIcon />,
+  VoltageGPU: <VoltageGPUIcon />,
   Vultr: <VultrIcon />,
   Wolfram: <WolframIcon />,
   XAI: <XAIIcon />,