NemoClaw User Configure Inference

Provider Options

The onboard wizard presents the following provider options by default. The first six are always available. Ollama appears when it is installed or running on the host.

Experimental Options

The following local inference options require NEMOCLAW_EXPERIMENTAL=1 and, when prerequisites are met, appear in the onboarding selection list.

For setup instructions, refer to Use a Local Inference Server (see the nemoclaw-user-configure-inference skill).

Validation

NemoClaw validates the selected provider and model before creating the sandbox. If validation fails, the wizard returns to provider selection.

Full details in references/inference-options.md.

Prerequisites

• A running NemoClaw sandbox.
• The OpenShell CLI on your PATH.
• NemoClaw installed.
• A local model server running, or Ollama installed. The NemoClaw onboard wizard can also start Ollama for you.

Change the active inference model while the sandbox is running. No restart is required.

Step 1: Switch to a Different Model

Switching happens through the OpenShell inference route. Use the provider and model that match the upstream you want to use.

NVIDIA Endpoints

$ openshell inference set --provider nvidia-prod --model nvidia/nemotron-3-super-120b-a12b

OpenAI

$ openshell inference set --provider openai-api --model gpt-5.4

Anthropic

$ openshell inference set --provider anthropic-prod --model claude-sonnet-4-6

Google Gemini

$ openshell inference set --provider gemini-api --model gemini-2.5-flash

Compatible Endpoints

If you onboarded a custom compatible endpoint, switch models with the provider created for that endpoint:

$ openshell inference set --provider compatible-endpoint --model <model-name>

$ openshell inference set --provider compatible-anthropic-endpoint --model <model-name>

If the provider itself needs to change, rerun nemoclaw onboard.

Switching from Responses API to Chat Completions

If onboarding selected /v1/responses but the agent fails at runtime (for example, because the backend does not emit the streaming events OpenClaw requires), re-run onboarding so the wizard re-probes the endpoint and bakes the correct API path into the image:

$ nemoclaw onboard

Select the same provider and endpoint again. The updated streaming probe will detect incomplete /v1/responses support and select /v1/chat/completions automatically.

For the compatible-endpoint provider, NemoClaw uses /v1/chat/completions by default, so no env var is required to keep the safe path. To opt in to /v1/responses for a backend you have verified end to end, set NEMOCLAW_PREFERRED_API before onboarding:

$ NEMOCLAW_PREFERRED_API=openai-responses nemoclaw onboard

Note: NEMOCLAW_INFERENCE_API_OVERRIDE patches the config at container startup but does not update the Dockerfile ARG baked into the image. If you recreate the sandbox without the override env var, the image reverts to the original API path. A fresh nemoclaw onboard is the reliable fix because it updates both the session and the baked image.

Step 2: Cross-Provider Switching

Switching to a different provider family (for example, from NVIDIA Endpoints to Anthropic) requires updating both the gateway route and the sandbox config.

Set the gateway route on the host:

$ openshell inference set --provider anthropic-prod --model claude-sonnet-4-6 --no-verify

Then set the override env vars and recreate the sandbox so they take effect at startup:

$ export NEMOCLAW_MODEL_OVERRIDE="anthropic/claude-sonnet-4-6"
$ export NEMOCLAW_INFERENCE_API_OVERRIDE="anthropic-messages"
$ nemoclaw onboard --resume --recreate-sandbox

The entrypoint patches openclaw.json at container startup with the override values. You do not need to rebuild the image. Remove the env vars and recreate the sandbox to revert to the original model.

NEMOCLAW_INFERENCE_API_OVERRIDE accepts openai-completions (for NVIDIA, OpenAI, Gemini, compatible endpoints) or anthropic-messages (for Anthropic and Anthropic-compatible endpoints). This variable is only needed when switching between provider families.

Step 3: Verify the Active Model

Run the status command to confirm the change:

$ nemoclaw <name> status

Add the --json flag for machine-readable output:

$ nemoclaw <name> status --json

The output includes the active provider, model, and endpoint.

Step 4: Notes

• The host keeps provider credentials.
• The sandbox continues to use inference.local.
• Same-provider model switches take effect immediately via the gateway route alone.
• Cross-provider switches also require NEMOCLAW_MODEL_OVERRIDE (and NEMOCLAW_INFERENCE_API_OVERRIDE) plus a sandbox recreate so the entrypoint patches the config at startup.
• Overrides are applied at container startup. Changing or removing env vars requires a sandbox recreate to take effect.

NemoClaw can route inference to a model server running on your machine instead of a cloud API. This page covers Ollama, compatible-endpoint paths for other servers, and two experimental options for vLLM and NVIDIA NIM.

All approaches use the same inference.local routing model. The agent inside the sandbox never connects to your model server directly. OpenShell intercepts inference traffic and forwards it to the local endpoint you configure.

Step 5: Ollama

Ollama is the default local inference option. The onboard wizard detects Ollama automatically when it is installed or running on the host.

If Ollama is not running, NemoClaw starts it for you. On macOS, the wizard also offers to install Ollama through Homebrew if it is not present.

Run the onboard wizard.

$ nemoclaw onboard

Select Local Ollama from the provider list. NemoClaw lists installed models or offers starter models if none are installed. It pulls the selected model, loads it into memory, and validates it before continuing.

Authenticated Reverse Proxy

NemoClaw keeps Ollama bound to 127.0.0.1:11434 and starts a token-gated reverse proxy on 0.0.0.0:11435. Containers and other hosts on the local network reach Ollama only through the proxy, which validates a Bearer token before forwarding requests. Ollama itself is never exposed without authentication.

The onboard wizard manages the proxy automatically:

• Generates a random 24-byte token on first run and stores it in ~/.nemoclaw/ollama-proxy-token with 0600 permissions.
• Starts the proxy after Ollama and verifies it before continuing.
• Cleans up stale proxy processes from previous runs.
• Reuses the persisted token after a host reboot so you do not need to re-run onboard.

The sandbox provider is configured to use proxy port 11435 with the generated token as its OPENAI_API_KEY credential. OpenShell's L7 proxy injects the token at egress, so the agent inside the sandbox never sees the token directly.

GET /api/tags is exempt from authentication so container health checks continue to work. All other endpoints (including POST /api/tags) require the Bearer token.

If Ollama is already running on a non-loopback address when you start onboard, the wizard restarts it on 127.0.0.1:11434 so the proxy is the only network path to the model server.

Non-Interactive Setup

$ NEMOCLAW_PROVIDER=ollama \
  NEMOCLAW_MODEL=qwen2.5:14b \
  nemoclaw onboard --non-interactive

If NEMOCLAW_MODEL is not set, NemoClaw selects a default model based on available memory.

Step 6: OpenAI-Compatible Server

This option works with any server that implements /v1/chat/completions, including vLLM, TensorRT-LLM, llama.cpp, LocalAI, and others. For compatible endpoints, NemoClaw uses /v1/chat/completions by default. This avoids a class of failures where local backends accept /v1/responses requests but silently drop the system prompt and tool definitions. To opt in to /v1/responses, set NEMOCLAW_PREFERRED_API=openai-responses before running onboard.

Start your model server. The examples below use vLLM, but any OpenAI-compatible server works.

$ vllm serve meta-llama/Llama-3.1-8B-Instruct --port 8000

Run the onboard wizard.

$ nemoclaw onboard

When the wizard asks you to choose an inference provider, select Other OpenAI-compatible endpoint. Enter the base URL of your local server, for example http://localhost:8000/v1.

The wizard prompts for an API key. If your server does not require authentication, enter any non-empty string (for example, dummy).

NemoClaw validates the endpoint by sending a test inference request before continuing. The wizard probes /v1/chat/completions by default for the compatible-endpoint provider. If you set NEMOCLAW_PREFERRED_API=openai-responses, NemoClaw probes /v1/responses instead and only selects it when the response includes the streaming events OpenClaw requires.

Non-Interactive Setup

Set the following environment variables for scripted or CI/CD deployments.

$ NEMOCLAW_PROVIDER=custom \
  NEMOCLAW_ENDPOINT_URL=http://localhost:8000/v1 \
  NEMOCLAW_MODEL=meta-llama/Llama-3.1-8B-Instruct \
  COMPATIBLE_API_KEY=dummy \
  nemoclaw onboard --non-interactive

Selecting the API Path

For the compatible-endpoint provider, /v1/chat/completions is the default. NemoClaw tests streaming events during onboarding and uses chat completions without probing the Responses API.

To opt in to /v1/responses, set NEMOCLAW_PREFERRED_API before running onboard:

$ NEMOCLAW_PREFERRED_API=openai-responses nemoclaw onboard

The wizard then probes /v1/responses and only selects it when streaming support is complete. If the probe fails, the wizard falls back to /v1/chat/completions automatically. You can use this variable in both interactive and non-interactive mode.

If you already onboarded and the sandbox is failing at runtime, re-run nemoclaw onboard to re-probe the endpoint and bake the correct API path into the image. Refer to Switch Inference Models (see the nemoclaw-user-configure-inference skill) for details.

Step 7: Anthropic-Compatible Server

If your local server implements the Anthropic Messages API (/v1/messages), choose Other Anthropic-compatible endpoint during onboarding instead.

$ nemoclaw onboard

For non-interactive setup, use NEMOCLAW_PROVIDER=anthropicCompatible and set COMPATIBLE_ANTHROPIC_API_KEY.

$ NEMOCLAW_PROVIDER=anthropicCompatible \
  NEMOCLAW_ENDPOINT_URL=http://localhost:8080 \
  NEMOCLAW_MODEL=my-model \
  COMPATIBLE_ANTHROPIC_API_KEY=dummy \
  nemoclaw onboard --non-interactive

Step 8: vLLM Auto-Detection (Experimental)

When vLLM is already running on localhost:8000, NemoClaw can detect it automatically and query the /v1/models endpoint to determine the loaded model.

Set the experimental flag and run onboard.

$ NEMOCLAW_EXPERIMENTAL=1 nemoclaw onboard

Select Local vLLM [experimental] from the provider list. NemoClaw detects the running model and validates the endpoint.

Note: NemoClaw forces the chat/completions API path for vLLM. The vLLM /v1/responses endpoint does not run the --tool-call-parser, so tool calls arrive as raw text.

Non-Interactive Setup

$ NEMOCLAW_EXPERIMENTAL=1 \
  NEMOCLAW_PROVIDER=vllm \
  nemoclaw onboard --non-interactive

NemoClaw auto-detects the model from the running vLLM instance. To override the model, set NEMOCLAW_MODEL.

Step 9: NVIDIA NIM (Experimental)

NemoClaw can pull, start, and manage a NIM container on hosts with a NIM-capable NVIDIA GPU.

Set the experimental flag and run onboard.

$ NEMOCLAW_EXPERIMENTAL=1 nemoclaw onboard

Select Local NVIDIA NIM [experimental] from the provider list. NemoClaw filters available models by GPU VRAM, pulls the NIM container image, starts it, and waits for it to become healthy before continuing.

Note: NIM uses vLLM internally. The same chat/completions API path restriction applies.

Non-Interactive Setup

$ NEMOCLAW_EXPERIMENTAL=1 \
  NEMOCLAW_PROVIDER=nim \
  nemoclaw onboard --non-interactive

To select a specific model, set NEMOCLAW_MODEL.

Step 10: Timeout Configuration

Local inference requests use a default timeout of 180 seconds. Large prompts on hardware such as DGX Spark can exceed shorter timeouts, so NemoClaw sets a higher default for local providers (Ollama, vLLM, NIM).

To override the timeout, set the NEMOCLAW_LOCAL_INFERENCE_TIMEOUT environment variable before onboarding:

$ export NEMOCLAW_LOCAL_INFERENCE_TIMEOUT=300
$ nemoclaw onboard

The value is in seconds. This setting is baked into the sandbox at build time. Changing it after onboarding requires re-running nemoclaw onboard.

Step 11: Verify the Configuration

After onboarding completes, confirm the active provider and model.

$ nemoclaw <name> status

The output shows the provider label (for example, "Local vLLM" or "Other OpenAI-compatible endpoint") and the active model.

Step 12: Switch Models at Runtime

You can change the model without re-running onboard. Refer to Switch Inference Models (see the nemoclaw-user-configure-inference skill) for the full procedure.

For compatible endpoints, the command is:

$ openshell inference set --provider compatible-endpoint --model <model-name>

If the provider itself needs to change (for example, switching from vLLM to a cloud API), rerun nemoclaw onboard.

Related Skills

• nemoclaw-user-get-started — Quickstart for first-time installation