IntegrationsHarness Integrations

Use the Remote Agent Adapter

View as Markdown

Use the nvidia.fabric.remote-agent adapter to call an agent service over HTTP or HTTPS. The service can implement OpenAI Responses, OpenAI Chat Completions, or Anthropic Messages.

Install the Adapter

Install the runtime and adapter together:

$pip install "nemo-fabric[remote-agent]"

To install the adapter without the runtime, use:

$pip install "nemo-fabric-adapters-remote-agent[harness]"

The bare, harness, and full installations contain the same adapter and HTTP client. They do not install the independently deployed remote service.

Configure the Adapter

Set the remote API root, including its /v1 path, in HarnessConfig.settings. api_type defaults to openai-responses.

1from nemo_fabric import HarnessConfig
2
3harness = HarnessConfig(
4    adapter_id="nvidia.fabric.remote-agent",
5    settings={
6        "base_url": "https://agent.example.com/v1",
7        "api_type": "openai-responses",
8        "connect_timeout_seconds": 10,
9        "read_timeout_seconds": 600,
10    },
11)

Supported api_type values are openai-responses, openai-completions, and anthropic-messages. Planning rejects unknown settings and values outside this set.

The adapter supports models, models.temperature, and replacement instructions.system values. Set models.default.api_key_env when the service requires a credential. The adapter sends it as a Bearer token for OpenAI APIs and as x-api-key for Anthropic Messages. For Anthropic Messages, you can set models.default.settings.max_tokens; when omitted, the adapter sends 4096. The connect timeout defaults to 10 seconds, and the timeout between response bytes defaults to 600 seconds.

The adapter exposes terminal invocation only. It does not support MCP, skills, tool policy, Relay telemetry, streaming, or subagents. It keeps the completed user/assistant transcript for ordered invocations within a runtime.