Release NeMo Relay 0.4.0 · NVIDIA/NeMo-Relay

We are proud to release NeMo Relay 0.4.0!

Key Features

🛡️ First-party PII redaction plugin with deterministic local backend support
🚦 Built-in NeMo Guardrails local backend support, so the nemo_guardrails component can run against either remote services or a Python-backed local worker.
⚙️ Code-driven plugin configuration layered over materialized global, project, and user plugin files while preserving documented precedence.
📡 Streaming ATOF export endpoints for HTTP POST, WebSocket, and long-lived NDJSON collectors.
🗄️ ATIF HTTP storage export support alongside S3-compatible trajectory storage.
💸 Pricing-aware LLM observability with model-pricing lookup, cost layering, and cost metadata propagated into ATIF, OpenInference, and OpenTelemetry output.
🧭 Stronger coding-agent trace fidelity across Hermes, OpenClaw, Claude Code, and Codex, including nested subagent lineage, routed-provider spans, and error-path consistency.
🧩 Improved framework integration behavior for LangChain, LangGraph, and Deep Agents, including serialization fixes, callback coverage, and model-response annotations.
🔌 Host plugin installation for Claude Code and Codex, including native plugin manifests, and install/uninstall/doctor --plugin subcommands.

chore(guardrails)!: remove python-only guardrails example plugin by @afourniernv in #198

feat: Improve Hermes Relay hook injection and event alignment by @yczhang-nv in #205
feat: add nested subagent session lineage for OpenClaw by @mnajafian-nv in #210
feat: add ATIF HTTP storage export by @willkill07 in #231
feat: add streaming ATOF endpoints by @willkill07 in #232
feat: add LLM cost layering and pricing lookup by @AjayThorve in #236
feat: add local backend for built-in nemo guardrails by @afourniernv in #197
feat: layer code-driven plugin config over materialized file configs by @zhongxuanwang-nv in #211
feat: add claude code and codex plugins by @willkill07 in #224
feat: add PII redaction plugin crate with local backend by @afourniernv in #245

fix: pass a real ScopeHandle to Node withScope callbacks by @zhongxuanwang-nv in #196
fix: improve OpenClaw observability consistency by @mnajafian-nv in #203
fix: remove race condition in LC integration tests by @dagardner-nv in #202
fix: improve OpenClaw observability cost consistency by @mnajafian-nv in #206
fix: add flattened OpenInference LLM attributes for annotations and replay by @mnajafian-nv in #207
fix: Fix LangChain input serialization in callback handler and fix deadlock in plugin context manager by @dagardner-nv in #208
fix: tighten OpenClaw ATOF hook-backed provenance metadata by @mnajafian-nv in #209
fix: suppress Claude Code lifecycle noise by @AjayThorve in #213
fix: preserve structured ATIF tool results by @dnandakumar-nv in #223
fix: validate Hermes routed provider observability by @mnajafian-nv in #222
fix: preserve Hermes error-path export consistency by @mnajafian-nv in #229
fix: preserve Hermes tool-result ATIF observations by @mnajafian-nv in #233
fix: resolve sanitized LLM request from annotations by @fallintoplace in #241
fix: avoid ATIF shutdown subscriber deadlock by @willkill07 in #242
fix: annotate Deep Agents model responses by @willkill07 in #246
fix: include pii redaction in version updates by @willkill07 in #251
fix: configure gateway body limits by @willkill07 in #261

test: validate OpenClaw nested subagent exports by @mnajafian-nv in #214
test: validate OpenClaw placeholder replay exports by @mnajafian-nv in #217
test: tighten Hermes wrapped observability contract by @mnajafian-nv in #215
test: validate OpenClaw model timing diagnostic exports by @mnajafian-nv in #218
test: cover Hermes wrapped ATIF fidelity and cost by @mnajafian-nv in #219
test: validate OpenClaw hook-only fallback exports by @mnajafian-nv in #220
test: add NeMo Relay skill eval datasets by @abhisawa-Nvidia in #225
test: validate Hermes routed provider OpenInference spans by @mnajafian-nv in #235
test: add Codex observability contract coverage by @yczhang-nv in #234
test: regenerate NeMo Relay evals by @abhisawa-Nvidia in #226
test: validate Hermes subagent lineage and orphan export coverage by @mnajafian-nv in #240
test(wasm): drop --test-force-exit to avoid Windows libuv abort by @zhongxuanwang-nv in #247
test: raise Rust and CLI coverage by @willkill07 in #259

docs: flush subscribers before deregister in package README examples by @zhongxuanwang-nv in #201
docs: register plugin kind before activating it in the plugin guide by @zhongxuanwang-nv in #200
docs: clarify Hermes integration paths by @mnajafian-nv in #249
docs: clarify exporter failure behavior by @willkill07 in #254
docs: pii redaction documentation by @afourniernv in #252
docs: update 0.4 release notes by @willkill07 in #262
docs: improve Relay onboarding flow by @willkill07 in #256