fix: update endpoint URLs to reflect new app naming conventions

modal/README.md

@@ -50,7 +50,7 @@ modal secret create huggingface-secret HF_TOKEN=hf_xxx   # for gated repos
 
 modal deploy modal/app_nvidia.py
 python modal/client.py \
-  --base-url https://<workspace>--nemotron-3-nano-4b.modal.run/v1 \
+  --base-url https://<workspace>--nvidia-llms-nemotron-3-nano-4b.modal.run/v1 \
   --model nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 \
   --prompt "Hello from the wood."
 ```

modal/app_nvidia.py

@@ -4,7 +4,7 @@ Deploy:    modal deploy modal/app_nvidia.py
 Serve dev: modal serve modal/app_nvidia.py
 
 Each model gets its own OpenAI-compatible endpoint at
-``https://<workspace>--<endpoint-name>.modal.run/v1`` (one per model in
+``https://<workspace>--nvidia-llms-<endpoint-name>.modal.run/v1`` (one per model in
 ``NVIDIA_MODELS``). Add or retune models in ``registry.py``.
 """
 

modal/client.py

@@ -2,10 +2,15 @@
 
 Usage:
     python modal/client.py \\
-        --base-url https://<workspace>--gemma-4-12b.modal.run/v1 \\
+        --base-url https://<workspace>--google-llms-gemma-4-12b.modal.run/v1 \\
         --model google/gemma-4-12B \\
         --prompt "Describe a mossy ticket booth in the wood."
 
+The endpoint URL is https://<workspace>--<app-name>-<endpoint-name>.modal.run/v1,
+where <app-name> is the modal.App (nvidia-llms / openbmb-llms / google-llms) and
+<endpoint-name> is the per-model slug.  --model is the served model id (the HF
+repo id), NOT the URL slug.
+
 The endpoints speak the OpenAI REST API, so the official ``openai`` SDK works
 unchanged — the engine can point ``OPENAI_BASE_URL`` at any of these URLs.
 """

modal/docs/deploying.md

@@ -38,19 +38,29 @@ shadow it).
 
 ## Endpoints
 
-Each model becomes its own OpenAI-compatible endpoint, named after its
-`endpoint_name`:
+Each model becomes its own OpenAI-compatible endpoint. Modal builds the URL from
+the `modal.App` name **and** the function's `endpoint_name`:
 
 ```
-https://<workspace>--<endpoint-name>.modal.run/v1
+https://<workspace>--<app-name>-<endpoint-name>.modal.run/v1
 ```
 
+`<app-name>` is `nvidia-llms`, `openbmb-llms`, or `google-llms` (one per provider
+app); `<endpoint-name>` is the per-model slug. e.g. the Nemotron 4B endpoint is
+`https://<workspace>--nvidia-llms-nemotron-3-nano-4b.modal.run/v1`.
+
+> **Model id vs URL slug.** The `--model` value (and the `"model"` field in a raw
+> request) is the *served model id* — the HF repo id, e.g.
+> `nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16` — because `served_model_name` defaults to
+> the repo `name`. It is **not** the URL slug (`nemotron-3-nano-4b`). Call
+> `/v1/models` on any endpoint to see the exact id it serves.
+
 Standard routes: `/v1/chat/completions`, `/v1/completions`, `/v1/models`, plus
 `/docs` for the Swagger UI. Smoke-test one:
 
 ```bash
 python modal/client.py \
-  --base-url https://<workspace>--gemma-4-12b.modal.run/v1 \
+  --base-url https://<workspace>--google-llms-gemma-4-12b.modal.run/v1 \
   --model google/gemma-4-12B \
   --prompt "Describe a mossy ticket booth in the wood."
 ```
@@ -130,7 +140,7 @@ Endpoints are OpenAI-compatible, so the engine talks to them through the OpenAI
 SDK. Point a model role at a deployed endpoint:
 
 ```bash
-export OPENAI_BASE_URL="https://<workspace>--nemotron-3-nano-4b.modal.run/v1"
+export OPENAI_BASE_URL="https://<workspace>--nvidia-llms-nemotron-3-nano-4b.modal.run/v1"
 export OPENAI_API_KEY="EMPTY"   # or the configured VLLM_API_KEY
 export MODEL_TINY="nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16"
 ```

modal/docs/openapi.md

@@ -8,8 +8,8 @@ the familiar OpenAI one. There are two sources of truth:
   `/docs`:
 
   ```
-  https://<workspace>--<endpoint-name>.modal.run/openapi.json
-  https://<workspace>--<endpoint-name>.modal.run/docs
+  https://<workspace>--<app-name>-<endpoint-name>.modal.run/openapi.json
+  https://<workspace>--<app-name>-<endpoint-name>.modal.run/docs
   ```
 
 - **Checked-in spec** — [`../openapi.yaml`](../openapi.yaml) documents the
@@ -19,11 +19,14 @@ the familiar OpenAI one. There are two sources of truth:
 ## Base URL
 
 ```
-https://<workspace>--<endpoint-name>.modal.run/v1
+https://<workspace>--<app-name>-<endpoint-name>.modal.run/v1
 ```
 
-One server per model; `<endpoint-name>` is the model's `endpoint_name` from
-`registry.py` (e.g. `gemma-4-12b`, `nemotron-3-nano-4b`).
+One server per model; the URL label is `<app-name>-<endpoint-name>` — the
+`modal.App` (`nvidia-llms` / `openbmb-llms` / `google-llms`) plus the model's
+`endpoint_name` from `registry.py` (e.g. `google-llms-gemma-4-12b`,
+`nvidia-llms-nemotron-3-nano-4b`). The `model` you send is the *served id* (the
+HF repo id), not this slug.
 
 ## Endpoints
 
@@ -60,7 +63,7 @@ otherwise. Clients pass the same token as their API key.
 ### curl
 
 ```bash
-curl https://<workspace>--gemma-4-12b.modal.run/v1/chat/completions \
+curl https://<workspace>--google-llms-gemma-4-12b.modal.run/v1/chat/completions \
   -H "Authorization: Bearer $LLM_API_KEY" \
   -H "Content-Type: application/json" \
   -d '{
@@ -76,7 +79,7 @@ curl https://<workspace>--gemma-4-12b.modal.run/v1/chat/completions \
 from openai import OpenAI
 
 client = OpenAI(
-    base_url="https://<workspace>--gemma-4-12b.modal.run/v1",
+    base_url="https://<workspace>--google-llms-gemma-4-12b.modal.run/v1",
     api_key=os.environ["LLM_API_KEY"],  # any value when auth is off
 )
 resp = client.chat.completions.create(
@@ -96,5 +99,5 @@ the `LLM_API_KEY` environment variable.
 openapi-generator-cli generate -i modal/openapi.yaml -g python -o ./gen
 
 # ...or from a live endpoint's exact spec:
-curl -s https://<workspace>--gemma-4-12b.modal.run/openapi.json -o openapi.json
+curl -s https://<workspace>--google-llms-gemma-4-12b.modal.run/openapi.json -o openapi.json
 ```

modal/openapi.yaml

@@ -19,15 +19,18 @@ servers:
         default: your-workspace
         description: Your Modal workspace slug.
       endpoint:
-        default: gemma-4-12b
-        description: A model's endpoint_name from registry.py.
+        default: google-llms-gemma-4-12b
+        description: >
+          The Modal function label: <app-name>-<endpoint_name>, where app-name is
+          the modal.App (nvidia-llms / openbmb-llms / google-llms) and endpoint_name
+          is the per-model slug from registry.py.
         enum:
-          - nemotron-3-nano-30b
-          - nemotron-3-nano-4b
-          - minicpm-o-4-5
-          - minicpm-4-1-8b
-          - gemma-4-26b
-          - gemma-4-12b
+          - nvidia-llms-nemotron-3-nano-30b
+          - nvidia-llms-nemotron-3-nano-4b
+          - openbmb-llms-minicpm-o-4-5
+          - openbmb-llms-minicpm-4-1-8b
+          - google-llms-gemma-4-26b
+          - google-llms-gemma-4-12b
 security:
   - bearerAuth: []
 paths: