nemo-relay-start

द्वारा nvidia

Help application developers pick a NeMo Relay binding and get to a first working scope, tool call, and LLM call

npx skills add https://github.com/nvidia/nemo-relay --skill nemo-relay-start

Get Started With NeMo Relay

Use this skill for first-time users who want the shortest path to a working example. Rust, Python, and Node.js are the primary quick-start and hosted-docs paths. Go and the raw FFI surface are source-first advanced paths.

Default Path

  • Pick the user's host language first: Rust, Python, or Node.js. If they are using Go or raw FFI, verify names against tracked source and tests.
  • Prefer the managed execution APIs over manual lifecycle APIs.
  • Start with one scope, one tool call, and one LLM call.
  • Add observability only after the basic flow works.

Guidance

  • Rust: use nemo_relay::api::scope::{push_scope, pop_scope, event} with builder params, then nemo_relay::api::tool::tool_call_execute(...) and nemo_relay::api::llm::llm_call_execute(...)
  • Python: uv sync, then use nemo_relay.scope.scope(...), nemo_relay.tools.execute(...), and nemo_relay.llm.execute(...)
  • Node.js: build the addon, then use withScope(...), toolCallExecute(...), and llmCallExecute(...)
  • Go: use source-first wrappers such as scope.Push(...), tools.Execute(...), llm.Execute(...), or top-level PushScope(...), ToolCallExecute(...), and LlmCallExecute(...)
  • FFI: recommend only for binding or embedding work; verify C names such as nemo_relay_push_scope, nemo_relay_tool_call_execute, and nemo_relay_llm_call_execute in the current header

Common Pitfalls

  • Calling execute APIs without an active scope
  • Skipping the build step for Rust, Python, or Node.js
  • Assuming source-first Go/FFI bindings have the same hosted-doc coverage as Rust, Python, and Node.js
  • Mixing manual lifecycle APIs into a first example

Embedded Quick-Start Notes

  • Install from packages when building a consumer app: Rust uses cargo add nemo-relay, Python uses uv add nemo-relay, and Node.js uses npm install nemo-relay-node.
  • Use repository setup commands when working from a checkout: Rust builds the workspace, Python rebuilds the virtual environment and native extension with uv sync, and Node.js installs and builds the native addon before tests or examples run.
  • A first example should register a short-lived subscriber, open an agent scope, emit one mark event, run one managed tool call, run one managed LLM call, then deregister the subscriber. In Python use nemo_relay.subscribers; in Node.js use root exports such as registerSubscriber; in Rust use nemo_relay::api::subscriber.
  • Success means the app emits scope start/end events plus tool and LLM lifecycle events, and the application result remains the provider or tool result.
  • Scope handles are explicit in Rust and optional in higher-level Python and Node.js helpers when the active scope is already correct. Pass the handle when the surrounding framework makes parentage ambiguous.

Related Skills

  • nemo-relay-instrument-calls
  • nemo-relay-setup-observability
  • nemo-relay-debug-runtime-integration

nvidia की और Skills

compileiq-debug
nvidia
Use when something is wrong: Search() hangs, all evaluations return INVALID_SCORE, scores aren't improving, every config returns the same number, ptxas errors…
official
create-github-pr
nvidia
Create GitHub pull requests using the gh CLI. Use when the user wants to create a new PR, submit code for review, or open a pull request. Trigger keywords -…
official
diagnose-perf
nvidia
First-responder performance triage for Isaac Sim and Isaac Lab. Identifies bottleneck category (GPU-bound, CPU-bound, VRAM, loading) using nvidia-smi and…
official
eagle3-review-logs
nvidia
Review EAGLE3 pipeline experiment logs from the launcher's experiments/ directory. Summarizes pass/fail status for all 4 tasks, diagnoses failures with root…
official
nemoclaw-maintainer-cross-issue-sweep
nvidia
Scans other open issues to find ones a given PR may also fix or accidentally break. Outputs adjacent-fix opportunities and contradiction risks with file:line…
official
karpathy-guidelines
nvidia
सामान्य LLM कोडिंग गलतियों को कम करने के लिए व्यवहार संबंधी दिशानिर्देश। कोड लिखते, समीक्षा करते या रिफैक्टर करते समय उपयोग करें ताकि अत्यधिक जटिलता से बचा जा सके, सर्जिकल बदलाव किए जा सकें,…
official
fhir-basics
nvidia
Teaches agents how FHIR R4 APIs work, what resources are available, how to query them with search parameters, and how to correctly parse all response formats…
official
underdeclared-agent
nvidia
A helpful assistant agent
official