Skip to content

Local Ollama configuration not reflected in sandbox openclaw.json (shows cloud nemotron default) #628

New issue
New issue
Open
#648
#630
Open
Local Ollama configuration not reflected in sandbox openclaw.json (shows cloud nemotron default)#628
#648
#630
Labels
bugSomething isn't workingstatus: triageFor new items that haven't been reviewed yet.
@AbirmanNvidia

Description

@AbirmanNvidia
AbirmanNvidia
opened on Mar 22, 2026

Description

Description

When selecting local Ollama during nemoclaw onboard, the sandbox's ~/.openclaw/openclaw.json still shows the cloud model (inference/nvidia/nemotron-3-super-120b-a12b) instead of the selected Ollama model.

Steps to Reproduce

  1. Run nemoclaw onboard --non-interactive with NEMOCLAW_PROVIDER=ollama and NEMOCLAW_MODEL=nemotron-3-nano:30b
  2. Or run interactive onboarding and select "Local Ollama" as the inference provider
  3. After completion, check ~/.openclaw/openclaw.json inside the sandbox

Expected Behavior

openclaw.json should reference the selected Ollama model (e.g., ollama/nemotron-3-nano:30b or similar).

Actual Behavior

openclaw.json contains:

"agents": {
  "defaults": {
    "model": {
      "primary": "inference/nvidia/nemotron-3-super-120b-a12b"
    }
  }
}

### Reproduction Steps



1) Install ollama, and pull LLM and embedding model for RAG


a. curl -fsSL https://ollama.com/install.sh | sh

b. ollama pull nemotron-3-nano:30b
We have single A100 GPU 80G on this machine, which can run nemotron-3-nano:30b, but not nemotron-3-super, it requires more GPU vRAM

root@host:~# lspci | grep -i nvidia
6b:00.0 3D controller: NVIDIA Corporation GA100 [A100 PCIe 80GB] (rev a1)
root@host:~#

c. ollama pull nomic-embed-text
nomic-embed-text is a text embedding model — it converts text into numerical vectors (embeddings) for semantic search, RAG, and similarity tasks.

ubuntu@host:~$ ollama list
NAME                       ID              SIZE      MODIFIED    
nomic-embed-text:latest    0a109f422b47    274 MB    2 hours ago    
nemotron-3-nano:30b        b725f1117407    24 GB     2 hours ago    
ubuntu@host:~$ 

ubuntu@host:~$ ollama list
NAME                       ID              SIZE      MODIFIED     
nomic-embed-text:latest    0a109f422b47    274 MB    11 hours ago    
nemotron-3-nano:30b        b725f1117407    24 GB     11 hours ago    
ubuntu@host:~$ curl http://localhost:11434/api/tags
{"models":[{"name":"nomic-embed-text:latest","model":"nomic-embed-text:latest","modified_at":"2026-03-21T10:27:44.641193503+02:00","size":274302450,"digest":"0a109f422b47e3a30ba2b10eca18548e944e8a23073ee3f3e947efcf3c45e59f","details":{"parent_model":"","format":"gguf","family":"nomic-bert","families":["nomic-bert"],"parameter_size":"137M","quantization_level":"F16"}},{"name":"nemotron-3-nano:30b","model":"nemotron-3-nano:30b","modified_at":"2026-03-21T10:16:03.326277546+02:00","size":24271934866,"digest":"b725f11174073334edd0c2ff396b8d4e66d7dab22a5a63717ccad5a08a270cf1","details":{"parent_model":"","format":"gguf","family":"nemotron_h_moe","families":["nemotron_h_moe"],"parameter_size":"31.6B","quantization_level":"Q4_K_M"}}]}ubuntu@host:~$ 

2) Install latest nodejs
curl -fsSL https://deb.nodesource.com/setup_22.x | sudo bash -
sudo apt-get install -y nodejs

ubuntu@host:~$ node --version
v22.22.1
ubuntu@host:~$ npm --version 
10.9.4
ubuntu@host:~$

3) Install Docker (required for sandboxes, podman is not supported at the moment)
curl -fsSL https://get.docker.com | sh
# Log out and back in for docker group to take effect

ubuntu@host:~$ id
uid=1000(ubuntu) gid=1000(ubuntu) groups=1000(ubuntu),4(adm),24(cdrom),27(sudo),30(dip),46(plugdev),101(lxd),987(docker)
ubuntu@host:~$

4) # Run the installer with sudo
curl -fsSL https://www.nvidia.com/nemoclaw.sh | sudo bash

######################
should local ollama be exposed on all interfaces?
######################
Created sandbox: my-assistant                                                                                                                                                                                                                                                                                                                                                     Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────
  Detected local inference option: Ollama
  Select one explicitly to use it. Press Enter to keep the cloud default.


  Inference options:
    1) NVIDIA Endpoint API (build.nvidia.com)
    2) Local Ollama (localhost:11434) — running (suggested)

  Choose [1]: 2
  ✓ Using Ollama on localhost:11434

  Ollama models:
    1) nomic-embed-text:latest
    2) nemotron-3-nano:30b

  Choose model [2]: 

  [5/7] Setting up inference provider
  ──────────────────────────────────────────────────
  Local Ollama is responding on localhost, but containers cannot reach http://host.openshell.internal:11434. Ensure Ollama listens on 0.0.0.0:11434 instead of 127.0.0.1 so sandboxes can reach it.
  On macOS, local inference also depends on OpenShell host routing support.
ubuntu@host:~$ 
###########################################

root@host:~# cat /etc/systemd/system/socat-ollama.service
[Unit]
Description=Forward Ollama to Docker bridge
After=network.target ollama.service
Requires=ollama.service

[Service]
Type=simple
ExecStart=/usr/bin/socat TCP-LISTEN:11434,bind=172.18.0.1,fork TCP:127.0.0.1:11434
Restart=always
RestartSec=5

[Install]
WantedBy=multi-user.target
root@host:~# 

listen on docker IP bridge, instead of all interfaces

root@host:~# curl http://172.18.0.1:11434/api/tags
{"models":[{"name":"nomic-embed-text:latest","model":"nomic-embed-text:latest","modified_at":"2026-03-21T10:27:44.641193503+02:00","size":274302450,"digest":"0a109f422b47e3a30ba2b10eca18548e944e8a23073ee3f3e947efcf3c45e59f","details":{"parent_model":"","format":"gguf","family":"nomic-bert","families":["nomic-bert"],"parameter_size":"137M","quantization_level":"F16"}},{"name":"nemotron-3-nano:30b","model":"nemotron-3-nano:30b","modified_at":"2026-03-21T10:16:03.326277546+02:00","size":24271934866,"digest":"b725f11174073334edd0c2ff396b8d4e66d7dab22a5a63717ccad5a08a270cf1","details":{"parent_model":"","format":"gguf","family":"nemotron_h_moe","families":["nemotron_h_moe"],"parameter_size":"31.6B","quantization_level":"Q4_K_M"}}]}root@host:~# 


host:~# nemoclaw --version
  Unknown command: --version

  Registered sandboxes: my-assistant
  Try: nemoclaw <sandbox-name> connect

  Run 'nemoclaw help' for usage.
root@host:~# nemoclaw my-assistant connect
sandbox@my-assistant:~$ 

#############################
Wrong provider is set, despite ollama was defined ealier
#############################

root@host:~# NEMOCLAW_RECREATE_SANDBOX=1 nemoclaw onboard --non-interactive

  NemoClaw Onboarding
  (non-interactive mode)
  ===================

  [1/7] Preflight checks
  ──────────────────────────────────────────────────
  ✓ Docker is running
  ✓ Container runtime: docker
  ✓ openshell CLI: openshell 0.0.13
  Cleaning up previous NemoClaw session...
  ✓ Previous session cleaned up
  ✓ Port 8080 available (OpenShell gateway)
  ✓ Port 18789 available (NemoClaw dashboard)
  ✓ NVIDIA GPU detected: 1 GPU(s), 81920 MB VRAM

  [2/7] Starting OpenShell gateway
  ──────────────────────────────────────────────────
  Using pinned OpenShell gateway image: ghcr.io/nvidia/openshell/cluster:0.0.13
✓ Checking Docker
✓ Downloading gateway
✓ Initializing environment
✓ Starting gateway                                                                                                                                                                       
✓ Gateway ready

  Name: nemoclaw
  Endpoint: https://127.0.0.1:8080

✓ Active gateway set to 'nemoclaw'
  ✓ Gateway is healthy

  [3/7] Creating sandbox
  ──────────────────────────────────────────────────
  [non-interactive] Sandbox name (lowercase, numbers, hyphens) [my-assistant]: → my-assistant
  [non-interactive] Sandbox 'my-assistant' exists — recreating
  Creating sandbox 'my-assistant' (this takes a few minutes on first run)...
Building image openshell/sandbox-from:1774124891 from /tmp/nemoclaw-build-0tGuw9/Dockerfile
  Context: /tmp/nemoclaw-build-0tGuw9
  Gateway: nemoclaw

  Building image openshell/sandbox-from:1774124891 from /tmp/nemoclaw-build-0tGuw9/Dockerfile
  Step 1/33 : FROM node:22-slim AS builder
   ---> 4f77a690f2f8
  Step 2/33 : COPY nemoclaw/package.json nemoclaw/tsconfig.json /opt/nemoclaw/
   ---> Using cache
   ---> 6293560dd804
  Step 3/33 : COPY nemoclaw/src/ /opt/nemoclaw/src/
   ---> Using cache
   ---> 8a8f53d4bb58
  Step 4/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 411021070ee7
  Step 5/33 : RUN npm install && npm run build
   ---> Using cache
   ---> 4319d89ae863
  Step 6/33 : FROM node:22-slim
   ---> 4f77a690f2f8
  Step 7/33 : ENV DEBIAN_FRONTEND=noninteractive
   ---> Using cache
   ---> 18e8a0e8cc76
  Step 8/33 : RUN apt-get update && apt-get install -y --no-install-recommends         python3 python3-pip python3-venv         curl git ca-certificates         iproute2     && rm -rf /var/lib/apt/lists/*
   ---> Using cache
   ---> 211cd92a3cec
  Step 9/33 : RUN groupadd -r sandbox && useradd -r -g sandbox -d /sandbox -s /bin/bash sandbox     && mkdir -p /sandbox/.nemoclaw     && chown -R sandbox:sandbox /sandbox
   ---> Using cache
   ---> eceb9d6e814f
  Step 10/33 : RUN mkdir -p /sandbox/.openclaw-data/agents/main/agent         /sandbox/.openclaw-data/extensions         /sandbox/.openclaw-data/workspace         /sandbox/.openclaw-data/skills         /sandbox/.openclaw-data/hooks         /sandbox/.openclaw-data/identity         /sandbox/.openclaw-data/devices         /sandbox/.openclaw-data/canvas         /sandbox/.openclaw-data/cron     && mkdir -p /sandbox/.openclaw     && ln -s /sandbox/.openclaw-data/agents /sandbox/.openclaw/agents     && ln -s /sandbox/.openclaw-data/extensions /sandbox/.openclaw/extensions     && ln -s /sandbox/.openclaw-data/workspace /sandbox/.openclaw/workspace     && ln -s /sandbox/.openclaw-data/skills /sandbox/.openclaw/skills     && ln -s /sandbox/.openclaw-data/hooks /sandbox/.openclaw/hooks     && ln -s /sandbox/.openclaw-data/identity /sandbox/.openclaw/identity     && ln -s /sandbox/.openclaw-data/devices /sandbox/.openclaw/devices     && ln -s /sandbox/.openclaw-data/canvas /sandbox/.openclaw/canvas     && ln -s /sandbox/.openclaw-data/cron /sandbox/.openclaw/cron     && touch /sandbox/.openclaw-data/update-check.json     && ln -s /sandbox/.openclaw-data/update-check.json /sandbox/.openclaw/update-check.json     && chown -R sandbox:sandbox /sandbox/.openclaw /sandbox/.openclaw-data
   ---> Using cache
   ---> 101147aa55f3
  Step 11/33 : RUN npm install -g openclaw@2026.3.11
   ---> Using cache
   ---> 1a9f1458190a
  Step 12/33 : RUN pip3 install --break-system-packages pyyaml
   ---> Using cache
   ---> 99d8c7f482a3
  Step 13/33 : COPY --from=builder /opt/nemoclaw/dist/ /opt/nemoclaw/dist/
   ---> Using cache
   ---> 6569631d0359
  Step 14/33 : COPY nemoclaw/openclaw.plugin.json /opt/nemoclaw/
   ---> Using cache
   ---> f18c7735977a
  Step 15/33 : COPY nemoclaw/package.json /opt/nemoclaw/
   ---> Using cache
   ---> 8d81a46ccf8e
  Step 16/33 : COPY nemoclaw-blueprint/ /opt/nemoclaw-blueprint/
   ---> Using cache
   ---> cba18e89956f
  Step 17/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 5e0956c494f2
  Step 18/33 : RUN npm install --omit=dev
   ---> Using cache
   ---> 085624626508
  Step 19/33 : RUN mkdir -p /sandbox/.nemoclaw/blueprints/0.1.0     && cp -r /opt/nemoclaw-blueprint/* /sandbox/.nemoclaw/blueprints/0.1.0/
   ---> Using cache
   ---> 69715a714c14
  Step 20/33 : COPY scripts/nemoclaw-start.sh /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> dee39ee4a8b1
  Step 21/33 : RUN chmod +x /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> e42747483a47
  Step 22/33 : ARG NEMOCLAW_MODEL=nvidia/nemotron-3-super-120b-a12b
   ---> Using cache
   ---> 351aa7d8efd7
  Step 23/33 : ARG CHAT_UI_URL=http://127.0.0.1:18789
   ---> Using cache
   ---> 314c4f152be2
  Step 24/33 : ARG NEMOCLAW_BUILD_ID=default
   ---> Using cache
   ---> cae95626e9ec
  Step 25/33 : WORKDIR /sandbox
   ---> Using cache
   ---> 4dd0905eaba8
  Step 26/33 : USER sandbox
   ---> Using cache
   ---> 83f051f2b0ef
  Step 27/33 : RUN python3 -c "import json, os, secrets; from urllib.parse import urlparse; model = '${NEMOCLAW_MODEL}'; chat_ui_url = '${CHAT_UI_URL}'; parsed = urlparse(chat_ui_url); chat_origin = f'{parsed.scheme}://{parsed.netloc}' if parsed.scheme and parsed.netloc else 'http://127.0.0.1:18789'; origins = ['http://127.0.0.1:18789']; origins = list(dict.fromkeys(origins + [chat_origin])); config = {     'agents': {'defaults': {'model': {'primary': f'inference/{model}'}}},     'models': {'mode': 'merge', 'providers': {         'nvidia': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'openshell-managed',             'api': 'openai-completions',             'models': [{'id': model.split('/')[-1], 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         },         'inference': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'unused',             'api': 'openai-completions',             'models': [{'id': model, 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         }     }},     'gateway': {         'mode': 'local',         'controlUi': {             'allowInsecureAuth': True,             'dangerouslyDisableDeviceAuth': True,             'allowedOrigins': origins,         },         'trustedProxies': ['127.0.0.1', '::1'],         'auth': {'token': secrets.token_hex(32)}     } }; path = os.path.expanduser('~/.openclaw/openclaw.json'); json.dump(config, open(path, 'w'), indent=2); os.chmod(path, 0o600)"
   ---> Using cache
   ---> 6c1fc48c0f7a
  Step 28/33 : RUN openclaw doctor --fix > /dev/null 2>&1 || true     && openclaw plugins install /opt/nemoclaw > /dev/null 2>&1 || true
   ---> Using cache
   ---> 28bc378760ef
  Step 29/33 : USER root
   ---> Using cache
   ---> feb0cc35ca9e
  Step 30/33 : RUN chown root:root /sandbox/.openclaw     && find /sandbox/.openclaw -mindepth 1 -maxdepth 1 -exec chown -h root:root {} +     && chmod 1777 /sandbox/.openclaw     && chmod 444 /sandbox/.openclaw/openclaw.json
   ---> Using cache
   ---> 9b05468d1f63
  Step 31/33 : USER sandbox
   ---> Using cache
   ---> bd1a78f66ab5
  Step 32/33 : ENTRYPOINT ["/bin/bash"]
   ---> Using cache
   ---> e154e6cd4b50
  Step 33/33 : CMD []
   ---> Using cache
   ---> 68ee53696acc
  Successfully built 68ee53696acc
  Successfully tagged openshell/sandbox-from:1774124891
  Built image openshell/sandbox-from:1774124891
  Pushing image openshell/sandbox-from:1774124891 into gateway "nemoclaw"
  [progress] Exported 1174 MiB
  [progress] Uploaded to gateway
  Image openshell/sandbox-from:1774124891 is available in the gateway.

✓ Image openshell/sandbox-from:1774124891 is available in the gateway.

                                                                                                                                                                                         Created sandbox: my-assistant                                                                                                                                                                                                                                                                                                                                                     Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────
  [non-interactive] Provider: cloud
  NVIDIA_API_KEY is required for cloud provider in non-interactive mode.
  Set it via: NVIDIA_API_KEY=nvapi-... nemoclaw onboard --non-interactive
root@host:~# 

####################################################################
#Second retry
####################################################################
root@host:~# NEMOCLAW_PROVIDER=ollama \
> NEMOCLAW_MODEL=nemotron-3-nano:30b \
> NEMOCLAW_RECREATE_SANDBOX=1 \
> nemoclaw onboard --non-interactive

  NemoClaw Onboarding
  (non-interactive mode)
  ===================

  [1/7] Preflight checks
  ──────────────────────────────────────────────────
  ✓ Docker is running
  ✓ Container runtime: docker
  ✓ openshell CLI: openshell 0.0.13
  Cleaning up previous NemoClaw session...
  ✓ Previous session cleaned up
  ✓ Port 8080 available (OpenShell gateway)
  ✓ Port 18789 available (NemoClaw dashboard)
  ✓ NVIDIA GPU detected: 1 GPU(s), 81920 MB VRAM

  [2/7] Starting OpenShell gateway
  ──────────────────────────────────────────────────
  Using pinned OpenShell gateway image: ghcr.io/nvidia/openshell/cluster:0.0.13
✓ Checking Docker
✓ Downloading gateway
✓ Initializing environment
✓ Starting gateway                                                                                                                                                                       
✓ Gateway ready

  Name: nemoclaw
  Endpoint: https://127.0.0.1:8080

✓ Active gateway set to 'nemoclaw'
  ✓ Gateway is healthy

  [3/7] Creating sandbox
  ──────────────────────────────────────────────────
  [non-interactive] Sandbox name (lowercase, numbers, hyphens) [my-assistant]: → my-assistant
  [non-interactive] Sandbox 'my-assistant' exists — recreating
  Creating sandbox 'my-assistant' (this takes a few minutes on first run)...
Building image openshell/sandbox-from:1774125448 from /tmp/nemoclaw-build-TuHrer/Dockerfile
  Context: /tmp/nemoclaw-build-TuHrer
  Gateway: nemoclaw

  Building image openshell/sandbox-from:1774125448 from /tmp/nemoclaw-build-TuHrer/Dockerfile
  Step 1/33 : FROM node:22-slim AS builder
   ---> 4f77a690f2f8
  Step 2/33 : COPY nemoclaw/package.json nemoclaw/tsconfig.json /opt/nemoclaw/
   ---> Using cache
   ---> 6293560dd804
  Step 3/33 : COPY nemoclaw/src/ /opt/nemoclaw/src/
   ---> Using cache
   ---> 8a8f53d4bb58
  Step 4/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 411021070ee7
  Step 5/33 : RUN npm install && npm run build
   ---> Using cache
   ---> 4319d89ae863
  Step 6/33 : FROM node:22-slim
   ---> 4f77a690f2f8
  Step 7/33 : ENV DEBIAN_FRONTEND=noninteractive
   ---> Using cache
   ---> 18e8a0e8cc76
  Step 8/33 : RUN apt-get update && apt-get install -y --no-install-recommends         python3 python3-pip python3-venv         curl git ca-certificates         iproute2     && rm -rf /var/lib/apt/lists/*
   ---> Using cache
   ---> 211cd92a3cec
  Step 9/33 : RUN groupadd -r sandbox && useradd -r -g sandbox -d /sandbox -s /bin/bash sandbox     && mkdir -p /sandbox/.nemoclaw     && chown -R sandbox:sandbox /sandbox
   ---> Using cache
   ---> eceb9d6e814f
  Step 10/33 : RUN mkdir -p /sandbox/.openclaw-data/agents/main/agent         /sandbox/.openclaw-data/extensions         /sandbox/.openclaw-data/workspace         /sandbox/.openclaw-data/skills         /sandbox/.openclaw-data/hooks         /sandbox/.openclaw-data/identity         /sandbox/.openclaw-data/devices         /sandbox/.openclaw-data/canvas         /sandbox/.openclaw-data/cron     && mkdir -p /sandbox/.openclaw     && ln -s /sandbox/.openclaw-data/agents /sandbox/.openclaw/agents     && ln -s /sandbox/.openclaw-data/extensions /sandbox/.openclaw/extensions     && ln -s /sandbox/.openclaw-data/workspace /sandbox/.openclaw/workspace     && ln -s /sandbox/.openclaw-data/skills /sandbox/.openclaw/skills     && ln -s /sandbox/.openclaw-data/hooks /sandbox/.openclaw/hooks     && ln -s /sandbox/.openclaw-data/identity /sandbox/.openclaw/identity     && ln -s /sandbox/.openclaw-data/devices /sandbox/.openclaw/devices     && ln -s /sandbox/.openclaw-data/canvas /sandbox/.openclaw/canvas     && ln -s /sandbox/.openclaw-data/cron /sandbox/.openclaw/cron     && touch /sandbox/.openclaw-data/update-check.json     && ln -s /sandbox/.openclaw-data/update-check.json /sandbox/.openclaw/update-check.json     && chown -R sandbox:sandbox /sandbox/.openclaw /sandbox/.openclaw-data
   ---> Using cache
   ---> 101147aa55f3
  Step 11/33 : RUN npm install -g openclaw@2026.3.11
   ---> Using cache
   ---> 1a9f1458190a
  Step 12/33 : RUN pip3 install --break-system-packages pyyaml
   ---> Using cache
   ---> 99d8c7f482a3
  Step 13/33 : COPY --from=builder /opt/nemoclaw/dist/ /opt/nemoclaw/dist/
   ---> Using cache
   ---> 6569631d0359
  Step 14/33 : COPY nemoclaw/openclaw.plugin.json /opt/nemoclaw/
   ---> Using cache
   ---> f18c7735977a
  Step 15/33 : COPY nemoclaw/package.json /opt/nemoclaw/
   ---> Using cache
   ---> 8d81a46ccf8e
  Step 16/33 : COPY nemoclaw-blueprint/ /opt/nemoclaw-blueprint/
   ---> Using cache
   ---> cba18e89956f
  Step 17/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 5e0956c494f2
  Step 18/33 : RUN npm install --omit=dev
   ---> Using cache
   ---> 085624626508
  Step 19/33 : RUN mkdir -p /sandbox/.nemoclaw/blueprints/0.1.0     && cp -r /opt/nemoclaw-blueprint/* /sandbox/.nemoclaw/blueprints/0.1.0/
   ---> Using cache
   ---> 69715a714c14
  Step 20/33 : COPY scripts/nemoclaw-start.sh /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> dee39ee4a8b1
  Step 21/33 : RUN chmod +x /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> e42747483a47
  Step 22/33 : ARG NEMOCLAW_MODEL=nvidia/nemotron-3-super-120b-a12b
   ---> Using cache
   ---> 351aa7d8efd7
  Step 23/33 : ARG CHAT_UI_URL=http://127.0.0.1:18789
   ---> Using cache
   ---> 314c4f152be2
  Step 24/33 : ARG NEMOCLAW_BUILD_ID=default
   ---> Using cache
   ---> cae95626e9ec
  Step 25/33 : WORKDIR /sandbox
   ---> Using cache
   ---> 4dd0905eaba8
  Step 26/33 : USER sandbox
   ---> Using cache
   ---> 83f051f2b0ef
  Step 27/33 : RUN python3 -c "import json, os, secrets; from urllib.parse import urlparse; model = '${NEMOCLAW_MODEL}'; chat_ui_url = '${CHAT_UI_URL}'; parsed = urlparse(chat_ui_url); chat_origin = f'{parsed.scheme}://{parsed.netloc}' if parsed.scheme and parsed.netloc else 'http://127.0.0.1:18789'; origins = ['http://127.0.0.1:18789']; origins = list(dict.fromkeys(origins + [chat_origin])); config = {     'agents': {'defaults': {'model': {'primary': f'inference/{model}'}}},     'models': {'mode': 'merge', 'providers': {         'nvidia': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'openshell-managed',             'api': 'openai-completions',             'models': [{'id': model.split('/')[-1], 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         },         'inference': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'unused',             'api': 'openai-completions',             'models': [{'id': model, 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         }     }},     'gateway': {         'mode': 'local',         'controlUi': {             'allowInsecureAuth': True,             'dangerouslyDisableDeviceAuth': True,             'allowedOrigins': origins,         },         'trustedProxies': ['127.0.0.1', '::1'],         'auth': {'token': secrets.token_hex(32)}     } }; path = os.path.expanduser('~/.openclaw/openclaw.json'); json.dump(config, open(path, 'w'), indent=2); os.chmod(path, 0o600)"
   ---> Using cache
   ---> 6c1fc48c0f7a
  Step 28/33 : RUN openclaw doctor --fix > /dev/null 2>&1 || true     && openclaw plugins install /opt/nemoclaw > /dev/null 2>&1 || true
   ---> Using cache
   ---> 28bc378760ef
  Step 29/33 : USER root
   ---> Using cache
   ---> feb0cc35ca9e
  Step 30/33 : RUN chown root:root /sandbox/.openclaw     && find /sandbox/.openclaw -mindepth 1 -maxdepth 1 -exec chown -h root:root {} +     && chmod 1777 /sandbox/.openclaw     && chmod 444 /sandbox/.openclaw/openclaw.json
   ---> Using cache
   ---> 9b05468d1f63
  Step 31/33 : USER sandbox
   ---> Using cache
   ---> bd1a78f66ab5
  Step 32/33 : ENTRYPOINT ["/bin/bash"]
   ---> Using cache
   ---> e154e6cd4b50
  Step 33/33 : CMD []
   ---> Using cache
   ---> 68ee53696acc
  Successfully built 68ee53696acc
  Successfully tagged openshell/sandbox-from:1774125448
  Built image openshell/sandbox-from:1774125448
  Pushing image openshell/sandbox-from:1774125448 into gateway "nemoclaw"
  [progress] Exported 1174 MiB
  [progress] Uploaded to gateway
  Image openshell/sandbox-from:1774125448 is available in the gateway.

✓ Image openshell/sandbox-from:1774125448 is available in the gateway.

                                                                                                                                                                                         Created sandbox: my-assistant                                                                                                                                                            
Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────
  [non-interactive] Provider: ollama
  ✓ Using Ollama on localhost:11434

  [5/7] Setting up inference provider
  ──────────────────────────────────────────────────
  Local Ollama is responding on localhost, but containers cannot reach http://host.openshell.internal:11434. Ensure Ollama listens on 0.0.0.0:11434 instead of 127.0.0.1 so sandboxes can reach it.
  On macOS, local inference also depends on OpenShell host routing support.
root@host:~# 

#######################################################################################
#Interactive install
#######################################################################################
Created sandbox: my-assistant                                                                                                                                                                                                                                                                                                                                                     Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────
  Detected local inference option: Ollama
  Select one explicitly to use it. Press Enter to keep the cloud default.


  Inference options:
    1) NVIDIA Endpoint API (build.nvidia.com)
    2) Local Ollama (localhost:11434) — running (suggested)

  Choose [1]: 

###############################################################
root@host:~# mkdir -p  /etc/systemd/system/ollama.service.d
root@host:~# vim /etc/systemd/system/ollama.service.d/override.conf
root@host:~# sudo systemctl daemon-reload
root@host:~# systemctl restart ollama
root@host:~# ss -tulpen | grep 11434
tcp   LISTEN 0      4096             172.18.0.1:11434      0.0.0.0:*    users:(("ollama",pid=592979,fd=3)) uid:996 ino:6450327 sk:3001 cgroup:/system.slice/ollama.service <->                       
root@host:~# nemoclaw onboard

  NemoClaw Onboarding
  ===================

  [1/7] Preflight checks
  ──────────────────────────────────────────────────
  ✓ Docker is running
  ✓ Container runtime: docker
  ✓ openshell CLI: openshell 0.0.13
  Cleaning up previous NemoClaw session...
  ✓ Previous session cleaned up
  ✓ Port 8080 available (OpenShell gateway)
  ✓ Port 18789 available (NemoClaw dashboard)
  ✓ NVIDIA GPU detected: 1 GPU(s), 81920 MB VRAM

  [2/7] Starting OpenShell gateway
  ──────────────────────────────────────────────────
  Using pinned OpenShell gateway image: ghcr.io/nvidia/openshell/cluster:0.0.13
✓ Checking Docker
✓ Downloading gateway
✓ Initializing environment
✓ Starting gateway                                                                                                                                                                       
✓ Gateway ready

  Name: nemoclaw
  Endpoint: https://127.0.0.1:8080

✓ Active gateway set to 'nemoclaw'
  ✓ Gateway is healthy

  [3/7] Creating sandbox
  ──────────────────────────────────────────────────
  Sandbox name (lowercase, numbers, hyphens) [my-assistant]: 
  Sandbox 'my-assistant' already exists. Recreate? [y/N]: y
  Creating sandbox 'my-assistant' (this takes a few minutes on first run)...
Building image openshell/sandbox-from:1774127592 from /tmp/nemoclaw-build-TrmijQ/Dockerfile
  Context: /tmp/nemoclaw-build-TrmijQ
  Gateway: nemoclaw

  Building image openshell/sandbox-from:1774127592 from /tmp/nemoclaw-build-TrmijQ/Dockerfile
  Step 1/33 : FROM node:22-slim AS builder
   ---> 4f77a690f2f8
  Step 2/33 : COPY nemoclaw/package.json nemoclaw/tsconfig.json /opt/nemoclaw/
   ---> Using cache
   ---> 6293560dd804
  Step 3/33 : COPY nemoclaw/src/ /opt/nemoclaw/src/
   ---> Using cache
   ---> 8a8f53d4bb58
  Step 4/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 411021070ee7
  Step 5/33 : RUN npm install && npm run build
   ---> Using cache
   ---> 4319d89ae863
  Step 6/33 : FROM node:22-slim
   ---> 4f77a690f2f8
  Step 7/33 : ENV DEBIAN_FRONTEND=noninteractive
   ---> Using cache
   ---> 18e8a0e8cc76
  Step 8/33 : RUN apt-get update && apt-get install -y --no-install-recommends         python3 python3-pip python3-venv         curl git ca-certificates         iproute2     && rm -rf /var/lib/apt/lists/*
   ---> Using cache
   ---> 211cd92a3cec
  Step 9/33 : RUN groupadd -r sandbox && useradd -r -g sandbox -d /sandbox -s /bin/bash sandbox     && mkdir -p /sandbox/.nemoclaw     && chown -R sandbox:sandbox /sandbox
   ---> Using cache
   ---> eceb9d6e814f
  Step 10/33 : RUN mkdir -p /sandbox/.openclaw-data/agents/main/agent         /sandbox/.openclaw-data/extensions         /sandbox/.openclaw-data/workspace         /sandbox/.openclaw-data/skills         /sandbox/.openclaw-data/hooks         /sandbox/.openclaw-data/identity         /sandbox/.openclaw-data/devices         /sandbox/.openclaw-data/canvas         /sandbox/.openclaw-data/cron     && mkdir -p /sandbox/.openclaw     && ln -s /sandbox/.openclaw-data/agents /sandbox/.openclaw/agents     && ln -s /sandbox/.openclaw-data/extensions /sandbox/.openclaw/extensions     && ln -s /sandbox/.openclaw-data/workspace /sandbox/.openclaw/workspace     && ln -s /sandbox/.openclaw-data/skills /sandbox/.openclaw/skills     && ln -s /sandbox/.openclaw-data/hooks /sandbox/.openclaw/hooks     && ln -s /sandbox/.openclaw-data/identity /sandbox/.openclaw/identity     && ln -s /sandbox/.openclaw-data/devices /sandbox/.openclaw/devices     && ln -s /sandbox/.openclaw-data/canvas /sandbox/.openclaw/canvas     && ln -s /sandbox/.openclaw-data/cron /sandbox/.openclaw/cron     && touch /sandbox/.openclaw-data/update-check.json     && ln -s /sandbox/.openclaw-data/update-check.json /sandbox/.openclaw/update-check.json     && chown -R sandbox:sandbox /sandbox/.openclaw /sandbox/.openclaw-data
   ---> Using cache
   ---> 101147aa55f3
  Step 11/33 : RUN npm install -g openclaw@2026.3.11
   ---> Using cache
   ---> 1a9f1458190a
  Step 12/33 : RUN pip3 install --break-system-packages pyyaml
   ---> Using cache
   ---> 99d8c7f482a3
  Step 13/33 : COPY --from=builder /opt/nemoclaw/dist/ /opt/nemoclaw/dist/
   ---> Using cache
   ---> 6569631d0359
  Step 14/33 : COPY nemoclaw/openclaw.plugin.json /opt/nemoclaw/
   ---> Using cache
   ---> f18c7735977a
  Step 15/33 : COPY nemoclaw/package.json /opt/nemoclaw/
   ---> Using cache
   ---> 8d81a46ccf8e
  Step 16/33 : COPY nemoclaw-blueprint/ /opt/nemoclaw-blueprint/
   ---> Using cache
   ---> cba18e89956f
  Step 17/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 5e0956c494f2
  Step 18/33 : RUN npm install --omit=dev
   ---> Using cache
   ---> 085624626508
  Step 19/33 : RUN mkdir -p /sandbox/.nemoclaw/blueprints/0.1.0     && cp -r /opt/nemoclaw-blueprint/* /sandbox/.nemoclaw/blueprints/0.1.0/
   ---> Using cache
   ---> 69715a714c14
  Step 20/33 : COPY scripts/nemoclaw-start.sh /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> dee39ee4a8b1
  Step 21/33 : RUN chmod +x /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> e42747483a47
  Step 22/33 : ARG NEMOCLAW_MODEL=nvidia/nemotron-3-super-120b-a12b
   ---> Using cache
   ---> 351aa7d8efd7
  Step 23/33 : ARG CHAT_UI_URL=http://127.0.0.1:18789
   ---> Using cache
   ---> 314c4f152be2
  Step 24/33 : ARG NEMOCLAW_BUILD_ID=default
   ---> Using cache
   ---> cae95626e9ec
  Step 25/33 : WORKDIR /sandbox
   ---> Using cache
   ---> 4dd0905eaba8
  Step 26/33 : USER sandbox
   ---> Using cache
   ---> 83f051f2b0ef
  Step 27/33 : RUN python3 -c "import json, os, secrets; from urllib.parse import urlparse; model = '${NEMOCLAW_MODEL}'; chat_ui_url = '${CHAT_UI_URL}'; parsed = urlparse(chat_ui_url); chat_origin = f'{parsed.scheme}://{parsed.netloc}' if parsed.scheme and parsed.netloc else 'http://127.0.0.1:18789'; origins = ['http://127.0.0.1:18789']; origins = list(dict.fromkeys(origins + [chat_origin])); config = {     'agents': {'defaults': {'model': {'primary': f'inference/{model}'}}},     'models': {'mode': 'merge', 'providers': {         'nvidia': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'openshell-managed',             'api': 'openai-completions',             'models': [{'id': model.split('/')[-1], 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         },         'inference': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'unused',             'api': 'openai-completions',             'models': [{'id': model, 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         }     }},     'gateway': {         'mode': 'local',         'controlUi': {             'allowInsecureAuth': True,             'dangerouslyDisableDeviceAuth': True,             'allowedOrigins': origins,         },         'trustedProxies': ['127.0.0.1', '::1'],         'auth': {'token': secrets.token_hex(32)}     } }; path = os.path.expanduser('~/.openclaw/openclaw.json'); json.dump(config, open(path, 'w'), indent=2); os.chmod(path, 0o600)"
   ---> Using cache
   ---> 6c1fc48c0f7a
  Step 28/33 : RUN openclaw doctor --fix > /dev/null 2>&1 || true     && openclaw plugins install /opt/nemoclaw > /dev/null 2>&1 || true
   ---> Using cache
   ---> 28bc378760ef
  Step 29/33 : USER root
   ---> Using cache
   ---> feb0cc35ca9e
  Step 30/33 : RUN chown root:root /sandbox/.openclaw     && find /sandbox/.openclaw -mindepth 1 -maxdepth 1 -exec chown -h root:root {} +     && chmod 1777 /sandbox/.openclaw     && chmod 444 /sandbox/.openclaw/openclaw.json
   ---> Using cache
   ---> 9b05468d1f63
  Step 31/33 : USER sandbox
   ---> Using cache
   ---> bd1a78f66ab5
  Step 32/33 : ENTRYPOINT ["/bin/bash"]
   ---> Using cache
   ---> e154e6cd4b50
  Step 33/33 : CMD []
   ---> Using cache
   ---> 68ee53696acc
  Successfully built 68ee53696acc
  Successfully tagged openshell/sandbox-from:1774127592
  Built image openshell/sandbox-from:1774127592
  Pushing image openshell/sandbox-from:1774127592 into gateway "nemoclaw"
  [progress] Exported 1174 MiB
  [progress] Uploaded to gateway
  Image openshell/sandbox-from:1774127592 is available in the gateway.

✓ Image openshell/sandbox-from:1774127592 is available in the gateway.

                                                                                                                                                                                         Created sandbox: my-assistant                                                                                                                                                            
Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────

  Inference options:
    1) NVIDIA Endpoint API (build.nvidia.com) (recommended)
    2) Local Ollama (localhost:11434)

  Choose [1]: 2
  Starting Ollama...
  ✓ Using Ollama on localhost:11434

  Ollama models:
    1) nemotron-3-nano:30b

  Choose model [1]: 

  [5/7] Setting up inference provider
  ──────────────────────────────────────────────────
  Local Ollama was selected, but nothing is responding on http://localhost:11434.
  On macOS, local inference also depends on OpenShell host routing support.
root@host:~# 

##################################################################################

The installation finishes only if ollama is listening on all interfaces

##################################################################################
root@host:~# nemoclaw onboard

  NemoClaw Onboarding
  ===================

  [1/7] Preflight checks
  ──────────────────────────────────────────────────
  ✓ Docker is running
  ✓ Container runtime: docker
  ✓ openshell CLI: openshell 0.0.13
  Cleaning up previous NemoClaw session...
  ✓ Previous session cleaned up
  ✓ Port 8080 available (OpenShell gateway)
  ✓ Port 18789 available (NemoClaw dashboard)
  ✓ NVIDIA GPU detected: 1 GPU(s), 81920 MB VRAM

  [2/7] Starting OpenShell gateway
  ──────────────────────────────────────────────────
  Using pinned OpenShell gateway image: ghcr.io/nvidia/openshell/cluster:0.0.13
✓ Checking Docker
✓ Downloading gateway
✓ Initializing environment
✓ Starting gateway                                                                                                                                                                       
✓ Gateway ready

  Name: nemoclaw
  Endpoint: https://127.0.0.1:8080

✓ Active gateway set to 'nemoclaw'
  ✓ Gateway is healthy

  [3/7] Creating sandbox
  ──────────────────────────────────────────────────
  Sandbox name (lowercase, numbers, hyphens) [my-assistant]: 
  Sandbox 'my-assistant' already exists. Recreate? [y/N]: y
  Creating sandbox 'my-assistant' (this takes a few minutes on first run)...
Building image openshell/sandbox-from:1774128091 from /tmp/nemoclaw-build-FcFxCy/Dockerfile
  Context: /tmp/nemoclaw-build-FcFxCy
  Gateway: nemoclaw

  Building image openshell/sandbox-from:1774128091 from /tmp/nemoclaw-build-FcFxCy/Dockerfile
  Step 1/33 : FROM node:22-slim AS builder
   ---> 4f77a690f2f8
  Step 2/33 : COPY nemoclaw/package.json nemoclaw/tsconfig.json /opt/nemoclaw/
   ---> Using cache
   ---> 6293560dd804
  Step 3/33 : COPY nemoclaw/src/ /opt/nemoclaw/src/
   ---> Using cache
   ---> 8a8f53d4bb58
  Step 4/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 411021070ee7
  Step 5/33 : RUN npm install && npm run build
   ---> Using cache
   ---> 4319d89ae863
  Step 6/33 : FROM node:22-slim
   ---> 4f77a690f2f8
  Step 7/33 : ENV DEBIAN_FRONTEND=noninteractive
   ---> Using cache
   ---> 18e8a0e8cc76
  Step 8/33 : RUN apt-get update && apt-get install -y --no-install-recommends         python3 python3-pip python3-venv         curl git ca-certificates         iproute2     && rm -rf /var/lib/apt/lists/*
   ---> Using cache
   ---> 211cd92a3cec
  Step 9/33 : RUN groupadd -r sandbox && useradd -r -g sandbox -d /sandbox -s /bin/bash sandbox     && mkdir -p /sandbox/.nemoclaw     && chown -R sandbox:sandbox /sandbox
   ---> Using cache
   ---> eceb9d6e814f
  Step 10/33 : RUN mkdir -p /sandbox/.openclaw-data/agents/main/agent         /sandbox/.openclaw-data/extensions         /sandbox/.openclaw-data/workspace         /sandbox/.openclaw-data/skills         /sandbox/.openclaw-data/hooks         /sandbox/.openclaw-data/identity         /sandbox/.openclaw-data/devices         /sandbox/.openclaw-data/canvas         /sandbox/.openclaw-data/cron     && mkdir -p /sandbox/.openclaw     && ln -s /sandbox/.openclaw-data/agents /sandbox/.openclaw/agents     && ln -s /sandbox/.openclaw-data/extensions /sandbox/.openclaw/extensions     && ln -s /sandbox/.openclaw-data/workspace /sandbox/.openclaw/workspace     && ln -s /sandbox/.openclaw-data/skills /sandbox/.openclaw/skills     && ln -s /sandbox/.openclaw-data/hooks /sandbox/.openclaw/hooks     && ln -s /sandbox/.openclaw-data/identity /sandbox/.openclaw/identity     && ln -s /sandbox/.openclaw-data/devices /sandbox/.openclaw/devices     && ln -s /sandbox/.openclaw-data/canvas /sandbox/.openclaw/canvas     && ln -s /sandbox/.openclaw-data/cron /sandbox/.openclaw/cron     && touch /sandbox/.openclaw-data/update-check.json     && ln -s /sandbox/.openclaw-data/update-check.json /sandbox/.openclaw/update-check.json     && chown -R sandbox:sandbox /sandbox/.openclaw /sandbox/.openclaw-data
   ---> Using cache
   ---> 101147aa55f3
  Step 11/33 : RUN npm install -g openclaw@2026.3.11
   ---> Using cache
   ---> 1a9f1458190a
  Step 12/33 : RUN pip3 install --break-system-packages pyyaml
   ---> Using cache
   ---> 99d8c7f482a3
  Step 13/33 : COPY --from=builder /opt/nemoclaw/dist/ /opt/nemoclaw/dist/
   ---> Using cache
   ---> 6569631d0359
  Step 14/33 : COPY nemoclaw/openclaw.plugin.json /opt/nemoclaw/
   ---> Using cache
   ---> f18c7735977a
  Step 15/33 : COPY nemoclaw/package.json /opt/nemoclaw/
   ---> Using cache
   ---> 8d81a46ccf8e
  Step 16/33 : COPY nemoclaw-blueprint/ /opt/nemoclaw-blueprint/
   ---> Using cache
   ---> cba18e89956f
  Step 17/33 : WORKDIR /opt/nemoclaw
   ---> Using cache
   ---> 5e0956c494f2
  Step 18/33 : RUN npm install --omit=dev
   ---> Using cache
   ---> 085624626508
  Step 19/33 : RUN mkdir -p /sandbox/.nemoclaw/blueprints/0.1.0     && cp -r /opt/nemoclaw-blueprint/* /sandbox/.nemoclaw/blueprints/0.1.0/
   ---> Using cache
   ---> 69715a714c14
  Step 20/33 : COPY scripts/nemoclaw-start.sh /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> dee39ee4a8b1
  Step 21/33 : RUN chmod +x /usr/local/bin/nemoclaw-start
   ---> Using cache
   ---> e42747483a47
  Step 22/33 : ARG NEMOCLAW_MODEL=nvidia/nemotron-3-super-120b-a12b
   ---> Using cache
   ---> 351aa7d8efd7
  Step 23/33 : ARG CHAT_UI_URL=http://127.0.0.1:18789
   ---> Using cache
   ---> 314c4f152be2
  Step 24/33 : ARG NEMOCLAW_BUILD_ID=default
   ---> Using cache
   ---> cae95626e9ec
  Step 25/33 : WORKDIR /sandbox
   ---> Using cache
   ---> 4dd0905eaba8
  Step 26/33 : USER sandbox
   ---> Using cache
   ---> 83f051f2b0ef
  Step 27/33 : RUN python3 -c "import json, os, secrets; from urllib.parse import urlparse; model = '${NEMOCLAW_MODEL}'; chat_ui_url = '${CHAT_UI_URL}'; parsed = urlparse(chat_ui_url); chat_origin = f'{parsed.scheme}://{parsed.netloc}' if parsed.scheme and parsed.netloc else 'http://127.0.0.1:18789'; origins = ['http://127.0.0.1:18789']; origins = list(dict.fromkeys(origins + [chat_origin])); config = {     'agents': {'defaults': {'model': {'primary': f'inference/{model}'}}},     'models': {'mode': 'merge', 'providers': {         'nvidia': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'openshell-managed',             'api': 'openai-completions',             'models': [{'id': model.split('/')[-1], 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         },         'inference': {             'baseUrl': 'https://inference.local/v1',             'apiKey': 'unused',             'api': 'openai-completions',             'models': [{'id': model, 'name': model, 'reasoning': False, 'input': ['text'], 'cost': {'input': 0, 'output': 0, 'cacheRead': 0, 'cacheWrite': 0}, 'contextWindow': 131072, 'maxTokens': 4096}]         }     }},     'gateway': {         'mode': 'local',         'controlUi': {             'allowInsecureAuth': True,             'dangerouslyDisableDeviceAuth': True,             'allowedOrigins': origins,         },         'trustedProxies': ['127.0.0.1', '::1'],         'auth': {'token': secrets.token_hex(32)}     } }; path = os.path.expanduser('~/.openclaw/openclaw.json'); json.dump(config, open(path, 'w'), indent=2); os.chmod(path, 0o600)"
   ---> Using cache
   ---> 6c1fc48c0f7a
  Step 28/33 : RUN openclaw doctor --fix > /dev/null 2>&1 || true     && openclaw plugins install /opt/nemoclaw > /dev/null 2>&1 || true
   ---> Using cache
   ---> 28bc378760ef
  Step 29/33 : USER root
   ---> Using cache
   ---> feb0cc35ca9e
  Step 30/33 : RUN chown root:root /sandbox/.openclaw     && find /sandbox/.openclaw -mindepth 1 -maxdepth 1 -exec chown -h root:root {} +     && chmod 1777 /sandbox/.openclaw     && chmod 444 /sandbox/.openclaw/openclaw.json
   ---> Using cache
   ---> 9b05468d1f63
  Step 31/33 : USER sandbox
   ---> Using cache
   ---> bd1a78f66ab5
  Step 32/33 : ENTRYPOINT ["/bin/bash"]
   ---> Using cache
   ---> e154e6cd4b50
  Step 33/33 : CMD []
   ---> Using cache
   ---> 68ee53696acc
  Successfully built 68ee53696acc
  Successfully tagged openshell/sandbox-from:1774128091
  Built image openshell/sandbox-from:1774128091
  Pushing image openshell/sandbox-from:1774128091 into gateway "nemoclaw"
  [progress] Exported 1174 MiB
  [progress] Uploaded to gateway
  Image openshell/sandbox-from:1774128091 is available in the gateway.

✓ Image openshell/sandbox-from:1774128091 is available in the gateway.

                                                                                                                                                                                         Created sandbox: my-assistant                                                                                                                                                                                                                                                                                                                                                     Setting up NemoClaw...
[gateway] openclaw gateway launched (pid 92)
[gateway] auto-pair watcher launched (pid 93)
[gateway] Local UI: http://127.0.0.1:18789/#token=<my token>
[gateway] Remote UI: http://127.0.0.1:18789/#token=<my token>
  Waiting for sandbox to become ready...
✓ Forwarding port 18789 to sandbox my-assistant in the background
  Access at: http://127.0.0.1:18789/
  Stop with: openshell forward stop 18789 my-assistant
  ✓ Sandbox 'my-assistant' created

  [4/7] Configuring inference (NIM)
  ──────────────────────────────────────────────────
  Detected local inference option: Ollama
  Select one explicitly to use it. Press Enter to keep the cloud default.


  Inference options:
    1) NVIDIA Endpoint API (build.nvidia.com)
    2) Local Ollama (localhost:11434) — running (suggested)

  Choose [1]: 2
  ✓ Using Ollama on localhost:11434

  Ollama models:
    1) nomic-embed-text:latest
    2) nemotron-3-nano:30b

  Choose model [2]: 2

  [5/7] Setting up inference provider
  ──────────────────────────────────────────────────
✓ Created provider ollama-local
Gateway inference configured:

  Route: inference.local
  Provider: ollama-local
  Model: nemotron-3-nano:30b
  Version: 1
  Priming Ollama model: nemotron-3-nano:30b
  ✓ Inference route set: ollama-local / nemotron-3-nano:30b

  [6/7] Setting up OpenClaw inside sandbox
  ──────────────────────────────────────────────────
  ✓ OpenClaw gateway launched inside sandbox

  [7/7] Policy presets
  ──────────────────────────────────────────────────

  Available policy presets:
    ○ discord — Discord API, gateway, and CDN access
    ○ docker — Docker Hub and NVIDIA container registry access
    ○ huggingface — Hugging Face Hub, LFS, and Inference API access
    ○ jira — Jira and Atlassian Cloud access
    ○ npm — npm and Yarn registry access (suggested)
    ○ outlook — Microsoft Outlook and Graph API access
    ○ pypi — Python Package Index (PyPI) access (suggested)
    ○ slack — Slack API and webhooks access
    ○ telegram — Telegram Bot API access

  Apply suggested presets (pypi, npm)? [Y/n/list]: list
  Enter preset names (comma-separated): 
  ✓ Policies applied

  ──────────────────────────────────────────────────
  Sandbox      my-assistant (Landlock + seccomp + netns)
  Model        nemotron-3-nano:30b (Local Ollama)
  NIM          not running
  ──────────────────────────────────────────────────
  Run:         nemoclaw my-assistant connect
  Status:      nemoclaw my-assistant status
  Logs:        nemoclaw my-assistant logs --follow
  ──────────────────────────────────────────────────

root@host:~# 

################################################################
In the chat. After configuring ollama to listen on 0.0.0.0:11434

what model is used?

The assistant is currently running the **Nemotron‑3‑Super** model (specifically `inference/nvidia/nemotron-3-super-120b-a12b`). This is also the default model for the session.

root@host:~# ollama list
NAME                       ID              SIZE      MODIFIED     
nomic-embed-text:latest    0a109f422b47    274 MB    13 hours ago    
nemotron-3-nano:30b        b725f1117407    24 GB     13 hours ago    
root@host:~#

################################################################

sandbox@my-assistant:~$ cat .openclaw/openclaw.json 
{
  "agents": {
    "defaults": {
      "model": {
        "primary": "inference/nvidia/nemotron-3-super-120b-a12b"
      }
    }
  },
  "models": {
    "mode": "merge",
    "providers": {
      "nvidia": {
        "baseUrl": "https://inference.local/v1",
        "apiKey": "openshell-managed",
        "api": "openai-completions",
        "models": [
          {
            "id": "nemotron-3-super-120b-a12b",
            "name": "nvidia/nemotron-3-super-120b-a12b",
            "reasoning": false,
            "input": [
              "text"
            ],
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 131072,
            "maxTokens": 4096
          }
        ]
      },
      "inference": {
        "baseUrl": "https://inference.local/v1",
        "apiKey": "unused",
        "api": "openai-completions",
        "models": [
          {
            "id": "nvidia/nemotron-3-super-120b-a12b",
            "name": "nvidia/nemotron-3-super-120b-a12b",
            "reasoning": false,
            "input": [
              "text"
            ],
            "cost": {
              "input": 0,
              "output": 0,
              "cacheRead": 0,
              "cacheWrite": 0
            },
            "contextWindow": 131072,
            "maxTokens": 4096
          }
        ]
      }
    }
  },
  "gateway": {
    "mode": "local",
    "controlUi": {
      "allowInsecureAuth": true,
      "dangerouslyDisableDeviceAuth": true,
      "allowedOrigins": [
        "http://127.0.0.1:18789"
      ]
    },
    "trustedProxies": [
      "127.0.0.1",
      "::1"
    ],
    "auth": {
      "token": "<my token>"
    }
  }
}sandbox@my-assistant:~$ 
###################################################################
root@host:~# openshell inference get
Gateway inference:

  Provider: ollama-local
  Model: nemotron-3-nano:30b
  Version: 1

System inference:

  Not configured
root@host:~#


### Environment

Environment

    NemoClaw: Latest (installed 2026-03-21)
    OpenShell: 0.0.13
    OS: Ubuntu 22.04
    GPU: A100 80GB

ubuntu@host:~$ node --version
v22.22.1
ubuntu@host:~$ npm --version 
10.9.4
ubuntu@host:~$
root@host:~# docker --version
Docker version 29.3.0, build 5927d80
root@host:~# 

[nemoclaw-debug.tar.gz](https://github.com/user-attachments/files/26164964/nemoclaw-debug.tar.gz)

### Debug Output

```shell

Logs



Checklist


    

  •  I confirmed this bug is reproducible
    • 

  •  I searched existing issues and this is not a duplicate
    • 

Metadata
Metadata
Assignees
No one assigned
    Labels
    bugSomething isn't workingstatus: triageFor new items that haven't been reviewed yet.
    Type
    No type
    Projects
    No projects
    Milestone
    No milestone
    Relationships
    None yet
    Development
    No branches or pull requests
    Issue actions