Skip to content

CHANGELOG.md

Latest commit

 

History

History
1490 lines (1434 loc) · 173 KB

CHANGELOG.md

File metadata and controls

1490 lines (1434 loc) · 173 KB
Raw

Changelog

All notable changes to this project will be documented in this file.

[1.7.0] - 2026-05-21

🚀 Notable Features and Improvements

  • Add AI coding agent skills for NeMo Agent Toolkit
  • Add consent-gated runtime telemetry for NeMo Agent Toolkit CLI commands

🚨 Breaking Changes

  • chore: migrate openai dependency from 1.x to 2.x by @willkill07 in NVIDIA#1849
  • chore: dependency cleanup - drops optuna chain and openinference from core by @bbednarski9 in NVIDIA#1902
  • Remove the nvidia-nat-vanna integration package by @dagardner-nv in NVIDIA#1926

✨ New Features

  • Add ATIF trajectory exporter for Phoenix visualization and debugging by @yczhang-nv in NVIDIA#1869
  • Add Exa Search API support as internet search tool by @maxwbuckley in NVIDIA#1846
  • Add OCI LangChain support for hosted Nemotron workflows by @fede-kamel in NVIDIA#1804
  • ATOF v0.1: Agentic Trajectory Observability Format (aligned spec) by @bbednarski9 in NVIDIA#1890
  • feat(cli): add consent-gated runtime telemetry for NAT CLI commands by @yczhang-nv in NVIDIA#1897
  • feat(observability): Arize AX OTLP exporter, docs, and examples by @ryoung562 in NVIDIA#1898
  • Feature: token streaming support for ReAct Agent by @thepatrickchin in NVIDIA#1851

🔧 Improvements

  • Misc CI improvements by @dagardner-nv in NVIDIA#1866
  • Misc CI improvements by @dagardner-nv in NVIDIA#1867
  • Support client-id-only MCP OAuth2 and refine Outlook auth example config by @AnuradhaKaruppiah in NVIDIA#1885
  • Misc test improvements and update model in the workflow template by @dagardner-nv in NVIDIA#1884
  • Widen the version range for starlette by @dagardner-nv in NVIDIA#1887
  • Filter AIMessage state updates from streaming output by @MylesShannon in NVIDIA#1863
  • Add optional --client-id and --client-secret flags to nat mcp client sub-commands by @dagardner-nv in NVIDIA#1891
  • dependency fix: remove flask as a hard dependency of nat-core by @bbednarski9 in NVIDIA#1895
  • dependency cleanup: move aioboto3, plotly, and wikipedia to dependent subpackages by @bbednarski9 in NVIDIA#1896
  • fix(tool): stop leaking exception details in chat_completion error response by @ColinM-sys in NVIDIA#1876
  • fix(example): Improve ADK-demo example with dynamic weather and time tools by @yczhang-nv in NVIDIA#1927
  • chore: bump dependency versions by @willkill07 in NVIDIA#1931
  • chore: bump dependency versions by @dagardner-nv in NVIDIA#1936
  • chore: Replace llama-3.2-nv-embedqa-1b-v2 with llama-nemotron-embed-1b-v2 by @dagardner-nv in NVIDIA#1937
  • chore: set dependency version range for azure-identity by @dagardner-nv in NVIDIA#1942
  • Replace llama-3.2-nv-rerankqa-1b-v2 with llama-nemotron-rerank-1b-v2 by @dagardner-nv in NVIDIA#1944
  • Update older models by @dagardner-nv in NVIDIA#1960
  • Adopt latest changes from NAT-UI by @dagardner-nv in NVIDIA#1963
  • Update model_health_check.py to check for deprecation header by @dagardner-nv in NVIDIA#1974

🐛 Bug Fixes

  • fix(memmachein): preserve conversation message order in add_items by @fede-kamel in NVIDIA#1856
  • Add an expiration time to cached MCP auth credentials by @dagardner-nv in NVIDIA#1872
  • Fix observability_evaluation_and_profiling example notebook by @dagardner-nv in NVIDIA#1874
  • Fix shutdown on Cntrl+c for ci/scripts/run_tests.py by @dagardner-nv in NVIDIA#1886
  • fix: preserve tool_calls in _stream_llm for native tool calling by @getglad in NVIDIA#1864
  • fix(mcp-client): set proper httpx timeouts for streamable-http transport by @yczhang-nv in NVIDIA#1889
  • fix(front-end): Fix HTTPInteractiveRunner streaming cleanup on client disconnect by @yczhang-nv in NVIDIA#1916
  • fix(deps): update selected dependency floors by @mnajafian-nv in NVIDIA#1923
  • fix(haystack): update urllib3 dependency floor by @mnajafian-nv in NVIDIA#1922
  • fix(agno): update Agno dependency stack by @mnajafian-nv in NVIDIA#1919
  • fix(fastmcp): update FastMCP to 3.2.4 by @mnajafian-nv in NVIDIA#1918
  • Remove currency A2A example from release validation by @AnuradhaKaruppiah in NVIDIA#1924
  • Update RL OpenPipe example by @dnandakumar-nv in NVIDIA#1925
  • fix(Phoenix): Fix Phoenix telemetry API key support by @yczhang-nv in NVIDIA#1928
  • fix(example): Fix typing.Any runtime checks in function conversion by @yczhang-nv in NVIDIA#1949
  • Fix MCP schema model reuse for Kaggle tools by @AnuradhaKaruppiah in NVIDIA#1954
  • fix(example): Fix auto-memory wrapper user ID lookup by @yczhang-nv in NVIDIA#1948
  • fix(mcp): Fix MCP reconnect lifecycle after local server crash by @yczhang-nv in NVIDIA#1935
  • Update LLM used in simple_auth_mcp example by @dagardner-nv in NVIDIA#1966
  • fix(example): Fix typing.Any and string input handling in NAT LangChain tools by @yczhang-nv in NVIDIA#1965
  • Fix auto memory recall scoping by @willkill07 in NVIDIA#1968
  • fix: improve auto memory wrapper example and code by @willkill07 in NVIDIA#1976
  • fix(server): emit canonical {"value": ...} envelope from /generate/full by @matthewgrossman in NVIDIA#1980

📝 Documentation Updates

  • Update LLMs used in the profiler documentation by @dagardner-nv in NVIDIA#1904
  • docs(memory): add Synap third-party memory plugin by @visy-ani in NVIDIA#1906
  • Add AI coding agent skills for NAT by @yczhang-nv in NVIDIA#1921
  • fix(skills): Fix skill-evolution routing for skill update decisions by @yczhang-nv in NVIDIA#1970
  • Update the Using Local LLMs documentation by @dagardner-nv in NVIDIA#1979

Other Changes

  • Forward-merge release/1.6 into develop by @rapids-bot[bot] in NVIDIA#1857
  • Forward-merge release/1.6 into develop by @rapids-bot[bot] in NVIDIA#1858
  • Forward-merge release/1.6 into develop by @rapids-bot[bot] in NVIDIA#1860

New Contributors

  • @fede-kamel made their first contribution in NVIDIA#1856
  • @maxwbuckley made their first contribution in NVIDIA#1846
  • @getglad made their first contribution in NVIDIA#1864
  • @ColinM-sys made their first contribution in NVIDIA#1876
  • @visy-ani made their first contribution in NVIDIA#1906
  • @ryoung562 made their first contribution in NVIDIA#1898
  • @matthewgrossman made their first contribution in NVIDIA#1980

[1.6.0] - 2026-04-09

🚀 Notable Features and Improvements

  • Adds ATIF-based evaluation, including a standalone evaluator class that can evaluate ATIF trajectories generated by any system

What's Changed

🚨 Breaking Changes

  • NAT Eval Dependency Reduction by @AnuradhaKaruppiah in NVIDIA#1690
  • Migrate nvidia_nat_ragas to Ragas v0.4 by @AnuradhaKaruppiah in NVIDIA#1780
  • Add support for optionally disabling SSL/TLS verification by @dagardner-nv in NVIDIA#1640
  • Add centralized user identity management and custom WebSocket authentication by @ericevans-nv in NVIDIA#1775
  • Feat/optimizer package extraction by @bhuvan002 in NVIDIA#1637
  • Decouple ATIF schemas from core and make nvidia-nat-eval lightweight by default by @AnuradhaKaruppiah in NVIDIA#1833

✨ New Features

  • Add TimeoutMiddleware by @ericevans-nv in NVIDIA#1677
  • feat(agent): stream tool call chunks by @MylesShannon in NVIDIA#1717
  • Add ATIF adapter: IntermediateStep to Harbor ATIF v1.6 trajectory conversion by @yczhang-nv in NVIDIA#1663
  • feat(control_flow): add built-in parallel_executor with example by @antoniomtz in NVIDIA#1733
  • Restore TimeoutMiddleware reverted by #1710 by @ericevans-nv in NVIDIA#1753
  • Add ATIF-native evaluator interface and runtime dual-lane dispatch by @AnuradhaKaruppiah in NVIDIA#1747
  • Add LLM error recovery, retry logic, and diagnostic logging to Tool Calling Agent by @ericevans-nv in NVIDIA#1757
  • Add ATIF-only custom evaluator example and docs by @AnuradhaKaruppiah in NVIDIA#1766
  • feat(profiler): ATIF-native profiler by @afourniernv in NVIDIA#1751
  • Add pre-tool verifier defense middleware for input instruction violation detection by @lidan-capsule in NVIDIA#1605
  • Add MemMachine memory integration for NeMo Agent Toolkit by @Charlie-Yi-2002 in NVIDIA#1460
  • Nested ATIF lineage by @AnuradhaKaruppiah in NVIDIA#1814

🔧 Improvements

  • Forward-merge release/1.5 into develop by @willkill07 in NVIDIA#1676
  • Add user attribution to Weave traces and feedback by @thepatrickchin in NVIDIA#1668
  • Move eval callbacks to nvidia_nat_eval and guard core shim path by @AnuradhaKaruppiah in NVIDIA#1748
  • Refactor eval file output into FileEvalCallback for Python API support by @yczhang-nv in NVIDIA#1743
  • Add AtifBaseEvaluator for reusable concurrent ATIF custom evaluators by @AnuradhaKaruppiah in NVIDIA#1769
  • Use nvidia-nat-eval from PyPI in the custom evaluator notebook by @AnuradhaKaruppiah in NVIDIA#1778
  • chore(pre-commit): any pyproject.toml changes should trigger uv lock checks by @willkill07 in NVIDIA#1790
  • Remove unneeded warning filter by @dagardner-nv in NVIDIA#1789
  • Remove old/unused compatibility code by @dagardner-nv in NVIDIA#1794
  • fix: handle Anthropic-style tool schemas in LangchainProfilerHandler by @cluster2600 in NVIDIA#1770
  • Add ATIF-native evaluators for nvidia-nat-profiler runtime metrics by @afourniernv in NVIDIA#1791
  • Update old CI variable names by @dagardner-nv in NVIDIA#1813
  • Explicitly set framework_wrappers for LangChain tools by @dagardner-nv in NVIDIA#1820
  • Report installed packages when testing built wheels by @dagardner-nv in NVIDIA#1821
  • chore: pin LiteLLM to known good version by @willkill07 in NVIDIA#1823
  • Ensure file paths are properly quoted in gitutils.py by @dagardner-nv in NVIDIA#1832
  • Update LICENSE-3rd-party.txt by @dagardner-nv in NVIDIA#1840
  • Informational ATIF mapping docs and sample output colocation (no functional changes) by @AnuradhaKaruppiah in NVIDIA#1844
  • fix: guard nim_langchain against ChatNVIDIA model lookup crash by @bbednarski9 in NVIDIA#1843
  • Document that MemMachine integration is Experimental by @AnuradhaKaruppiah in NVIDIA#1854
  • Restore profiler intermediate-step compatibility by rolling back selected ATIF-native changes by @AnuradhaKaruppiah in NVIDIA#1853

🐛 Bug Fixes

  • Fix Simple Calculator MCP example to return consistent timezone results by @dagardner-nv in NVIDIA#1708
  • fix: preserve nested Pydantic models when unpacking multi-arg functions by @MylesShannon in NVIDIA#1696
  • Fix _running_workflow_task always being None due to method chaining by @AjayThorve in NVIDIA#1744
  • Fix LoRA inference routing in OpenPipe ART cookbook by @aslanshi in NVIDIA#1662
  • Add profiler callback to sequential executor for LLM observability by @thepatrickchin in NVIDIA#1786
  • Enhance chat completion functionality to support OpenAI-style message history by @mdemoret-nv in NVIDIA#1674
  • fix(crewai): CrewAI Profiler Handler logic >=1.1.0 by @willkill07 in NVIDIA#1803
  • fix: add missing dependencies and repair logic for some tests by @willkill07 in NVIDIA#1815
  • Fix urls in pypi markdown files by @dagardner-nv in NVIDIA#1828
  • Fix default values for datetime fields in JobInfo objects by @dagardner-nv in NVIDIA#1829
  • add targeted MCP OAuth2 diagnostics for redirect URI mismatches by @AnuradhaKaruppiah in NVIDIA#1830
  • Ensure proper cancellation of in-flight workflow tasks by @thepatrickchin in NVIDIA#1819
  • Ensure HTTP request attributes are propagated into eval and async requests by @dagardner-nv in NVIDIA#1838
  • Update weave version to pickup fixes for tracing LangChain workflows by @dagardner-nv in NVIDIA#1841
  • fix(middleware): improve pre-tool middleware guarding logic by @cparadis-nvidia in NVIDIA#1824

📝 Documentation Updates

  • Add notebooks demonstrating standalone ATIF evaluation via Python API by @yczhang-nv in NVIDIA#1752
  • Add notebook demonstrating end-to-end interoperability between third-party agent frameworks and the NeMo Agent Toolkit evaluation harness by @yczhang-nv in NVIDIA#1799
  • docs: add LangSmith to list of supported observability platforms by @abhisawa-Nvidia in NVIDIA#1825
  • Fix license header in memmachine's README.md by @Charlie-Yi-2002 in NVIDIA#1831
  • add memmachine into docs/ by @Charlie-Yi-2002 in NVIDIA#1839
  • Document Otel endpoints for custom LangSmith deployments by @dagardner-nv in NVIDIA#1845

New Contributors

  • @AjayThorve made their first contribution in NVIDIA#1744
  • @cluster2600 made their first contribution in NVIDIA#1770
  • @afourniernv made their first contribution in NVIDIA#1791
  • @bhuvan002 made their first contribution in NVIDIA#1637
  • @lidan-capsule made their first contribution in NVIDIA#1605
  • @abhisawa-Nvidia made their first contribution in NVIDIA#1825
  • @Charlie-Yi-2002 made their first contribution in NVIDIA#1460
  • @cparadis-nvidia made their first contribution in NVIDIA#1824

[1.5.0] - 2026-03-11

🚀 Notable Features and Improvements

  • Dynamo Runtime Intelligence: Automatically infer per-request latency sensitivity from agent profiles and apply runtime hints for cache control, load-aware routing, and priority-aware serving.
  • Agent Performance Primitives (APP): Introduce framework-agnostic performance primitives that accelerate graph-based agent frameworks such as LangChain, CrewAI, and Agno with parallel execution, speculative branching, and node-level priority routing.
  • LangSmith Native Integration: Observe end-to-end agent execution with native LangSmith tracing, run evaluation experiments, compare outcomes, and manage prompt versions across development and production workflows.
  • FastMCP Workflow Publishing: Publish NeMo Agent Toolkit workflows as MCP servers using the FastMCP runtime to simplify MCP-native deployment and integration.

Migration notice: 1.5.0 includes packaging and compatibility refactors (meta-package restructure, eval/profiling package split, and import-path updates). See the Migration Guide.

What's Changed

🚨 Breaking Changes

  • Dynamic Inference Headers with Prediction Trie Integration by @dnandakumar-nv in NVIDIA#1483
  • improvement(packaging): Packaging Restructure for libraries by @willkill07 in NVIDIA#1512
  • fix: Langchain agents should reuse runnable config by @willkill07 in NVIDIA#1604
  • Refactor: Split eval/profiler into optional nvidia-nat-eval package by @AnuradhaKaruppiah in NVIDIA#1599
  • improvement: nvext.agent_hints and nvext.cache_control clean up by @bbednarski9 in NVIDIA#1648

✨ New Features

  • Add evaluator feedback to GA optimizer by @dnandakumar-nv in NVIDIA#1442
  • Add similarity scores and threshold filtering to Redis semantic search by @thepatrickchin in NVIDIA#1434
  • Add end-to-end custom metadata propagation for observability by @ericevans-nv in NVIDIA#1480
  • Expose MCP session ID and add custom headers support for session correlation by @yczhang-nv in NVIDIA#1500
  • Preserve workflow state across WebSocket reconnections by @ericevans-nv in NVIDIA#1541
  • feat(ci-scripts): utility scripts for license updates and SBOM by @willkill07 in NVIDIA#1548
  • Add example of control flow composition with router agent and sequential executor by @thepatrickchin in NVIDIA#1550
  • Add support for publishing a NeMo Agent Toolkit workflow as a MCP server via FastMCP3 by @AnuradhaKaruppiah in NVIDIA#1539
  • Add dataset store registration and support by @dnandakumar-nv in NVIDIA#1576
  • Allow for prompts to be stored in/loaded from files by @pastorsj in NVIDIA#1567
  • feat(observability): Cross-Workflow Observability by @willkill07 in NVIDIA#1598
  • FastAPI Frontend Refactor with HTTP HITL and OAuth Support by @willkill07 in NVIDIA#1603
  • Add support for LangSmith evaluators by @mpenn in NVIDIA#1592
  • Add automatic latency sensitivity inference by @dnandakumar-nv in NVIDIA#1618
  • RAG Library Mode integration by @ericevans-nv in NVIDIA#1440
  • feat: Add HuggingFace Inference API and Embedder providers by @bledden in NVIDIA#1570
  • Integrate LangSmith Observability with Evaluation and Optimization by @pastorsj in NVIDIA#1593
  • Add experimental nvidia-nat-app Agent Performance Primitives subpackage by @mpenn in NVIDIA#1636
  • Add Dynamo Example for Latency Sensitivity Assignment by @dnandakumar-nv in NVIDIA#1634
  • Revert unintended changes from PR #1704, preserve nat-ui submodule fix by @ericevans-nv in NVIDIA#1710

🔧 Improvements

  • Restore version 1.5 on develop after forward merge by @mnajafian-nv in NVIDIA#1324
  • Forward-merge release/1.4 into develop (conflict resolution) by @mnajafian-nv in NVIDIA#1394
  • Add OAuth2-Protected MCP Calculator Example by @AnuradhaKaruppiah in NVIDIA#1403
  • Forward-merge release/1.4 into develop by @mnajafian-nv in NVIDIA#1453
  • Merge release/1.4 into develop by @willkill07 in NVIDIA#1459
  • Add use_native_tool_calling option to ReAct agent by @yczhang-nv in NVIDIA#1476
  • Add raise_on_parsing_failure option to ReAct agent by @yczhang-nv in NVIDIA#1477
  • Enable per-user workflow support in nat eval by @ericevans-nv in NVIDIA#1503
  • feat: make tavily internet search tool configurable by @cdgamarose-nv in NVIDIA#1518
  • Update nat-ui submodule to latest main by @ericevans-nv in NVIDIA#1551
  • chore(pre-commit): update versions in pre-commit; add root-level uv.lock check by @willkill07 in NVIDIA#1553
  • Add a new per-user MCP client tool list endpoint by @AnuradhaKaruppiah in NVIDIA#1561
  • Add HTTP retry logic and error resilience for workflow evaluation by @ericevans-nv in NVIDIA#1563
  • Fix workflow name regression introduced by custom OTEL span naming by @ericevans-nv in NVIDIA#1572
  • Add Support for NVExt Annotations and Latency Sensitivity for Dynamo by @dnandakumar-nv in NVIDIA#1575
  • chore(ci): add stale action for old Issues/PRs by @willkill07 in NVIDIA#1581
  • Update dynamo headers to provide raw integer values by default by @dnandakumar-nv in NVIDIA#1583
  • Resolve user ID from JWT or nat-session cookie by @AnuradhaKaruppiah in NVIDIA#1584
  • Add support for Weave feedback comments by @thepatrickchin in NVIDIA#1586
  • chore(llm-providers): Add env OPENAI_BASE_URL for openai ; unify llm provider configs by @willkill07 in NVIDIA#1577
  • HITL prompt timeouts and API error responses by @ericevans-nv in NVIDIA#1591
  • Update nat-ui submodule to latest main by @ericevans-nv in NVIDIA#1594
  • enh(sbom-licenses): refactor common code; enable multi-version diffs by @willkill07 in NVIDIA#1597
  • Populate full connection attributes and payload for HTTP and WebSocket sessions by @ericevans-nv in NVIDIA#1602
  • Refactor latency sensitivity to use integers instead of enums. by @dnandakumar-nv in NVIDIA#1601
  • chore: update nat-ui submodule by @ericevans-nv in NVIDIA#1606
  • Add cache pinning strategy for KV cache with TTL control by @dnandakumar-nv in NVIDIA#1609
  • Add type converters for langgraph wrapper nat serve endpoints by @ericevans-nv in NVIDIA#1610
  • Add max_sensitivity for latency-based prioritization by @dnandakumar-nv in NVIDIA#1612
  • feat(agent): add token-by-token streaming to tool_calling_agent by @MylesShannon in NVIDIA#1595
  • chore(deps): upgrade uv.lock deps prior to release; relax dependencies by @willkill07 in NVIDIA#1621
  • Allow running pytest from project root by @dagardner-nv in NVIDIA#1622
  • Fix warning messages emitted from test_per_user_fastapi_integration.py by @dagardner-nv in NVIDIA#1624
  • feature: dynamo integration with nat profiler and prometheus/grafana dashboard by @bbednarski9 in NVIDIA#1486
  • Remove ci/release/update_toml_dep.py script by @dagardner-nv in NVIDIA#1646
  • CI fix: exclude CHANGELOG.md from pre-commit checks by @bbednarski9 in NVIDIA#1653
  • Remove NASSE naming by @ericevans-nv in NVIDIA#1632
  • chore: remove all unnecessary docker deployment guides from examples by @willkill07 in NVIDIA#1655
  • Remove Profiler Agent Example from the Primary Toolkit Repo by @AnuradhaKaruppiah in NVIDIA#1656
  • improvement: nvext.cache_control warning and HiCache for SgLang images by @bbednarski9 in NVIDIA#1658
  • Fix multi_frameworks example UnboundLocalError and upgrade default LLM by @ericevans-nv in NVIDIA#1661
  • Fixes and improvements for tests by @dagardner-nv in NVIDIA#1659
  • improvement(logging): add file logging mode option by @willkill07 in NVIDIA#1651
  • Refactor defense and red teaming middleware with pre/post invoke hooks by @ericevans-nv in NVIDIA#1671
  • Add a server-side override for the A2A Agent Card URL by @AnuradhaKaruppiah in NVIDIA#1673
  • Fix Mem0 metadata validation error and improve auto_memory_wrapper example by @ericevans-nv in NVIDIA#1683
  • Add an optional proxy server to map model names for integration testing by @dagardner-nv in NVIDIA#1679
  • chore(deps): bump package versions by @willkill07 in NVIDIA#1682
  • Fix DynamicFunctionMiddleware builder patching regression by @ericevans-nv in NVIDIA#1691
  • Fix incorrect CLI flag in auto_memory_wrapper README by @ericevans-nv in NVIDIA#1692
  • Add E2E test for Tool Calling Responses API Agent by @dagardner-nv in NVIDIA#1726
  • Fix Milvus connection failures in RAG integration tests by @ericevans-nv in NVIDIA#1724
  • feat: add NIM model endpoint health check by @mnajafian-nv in NVIDIA#1716
  • feat: add embedder inference check and Slack reporting for model health by @mnajafian-nv in NVIDIA#1736
  • Update container tag for the nginx-rewrite-models service by @dagardner-nv in NVIDIA#1740
  • Improve the text file ingest E2E test by @dagardner-nv in NVIDIA#1759
  • Update uv.lock files by @dagardner-nv in NVIDIA#1762
  • Improves FastMCP dev experience and docs by @AnuradhaKaruppiah in NVIDIA#1773
  • Observability user experiences fixes by @dnandakumar-nv in NVIDIA#1760

🐛 Bug Fixes

  • fix: bump NAT version to 1.5 for packages that were added under release/1.4 by @willkill07 in NVIDIA#1399
  • Fix MCP tool validation for nullable optional fields by @AnuradhaKaruppiah in NVIDIA#1507
  • fix(serve): ensure a single event loop for python 3.11 by @willkill07 in NVIDIA#1528
  • fix: flaky batching processor test by @willkill07 in NVIDIA#1529
  • fix(ci): coverage reports should only be for nat code and examples by @willkill07 in NVIDIA#1536
  • fix(ci): Fix build_wheels and slack notifications for nightlies by @willkill07 in NVIDIA#1537
  • fix(tests): add required deps for some e2e tests; get notebook tests working by @willkill07 in NVIDIA#1538
  • Forward-merge release/1.4 into develop by @willkill07 in NVIDIA#1552
  • Use relative paths for symlink creation in workflow create command by @thepatrickchin in NVIDIA#1557
  • fix(milvus): Fix vector_field config mapping and document_id type by @rmalani-nv in NVIDIA#1555
  • Refactor span attribute serialization to use JSON strings by @dnandakumar-nv in NVIDIA#1574
  • fix(ci): ensure packaging works in GitLab CI by @willkill07 in NVIDIA#1582
  • Fix FastMCP example E2E tests by @AnuradhaKaruppiah in NVIDIA#1580
  • fix(gitlab-ci): ensure gitlab artifact upload is configured correctly by @willkill07 in NVIDIA#1588
  • fix(ci): ensure stale action has required permissions by @willkill07 in NVIDIA#1589
  • fix(mcp): Cache enum classes to prevent validation errors by @bledden in NVIDIA#1564
  • fix(tests): prepare for OpenAI endpoint for nightlies; fix failing tests by @willkill07 in NVIDIA#1596
  • Refactor call index tracking for prefix predictions by @dnandakumar-nv in NVIDIA#1608
  • Fix failures after nvidia-nat-eval isolation by @willkill07 in NVIDIA#1615
  • fix(mcp-client): ensure tools are only invoked when available by @willkill07 in NVIDIA#1616
  • fix: update openpipe-art accuracy reward logic by @aslanshi in NVIDIA#1623
  • fix: Preserve custom dataset fields in workflow output by @bledden in NVIDIA#1628
  • fix: Skip output directory cleanup when --skip_workflow is set by @bledden in NVIDIA#1627
  • fix: Pass request_timeout through to OpenAI/Azure LLM clients by @bledden in NVIDIA#1626
  • fix: Filter empty LLM responses from ReAct retry scratchpad by @bledden in NVIDIA#1629
  • Fix auth callback trace and update test scripts by @AnuradhaKaruppiah in NVIDIA#1633
  • fix(http-hitl-oauth): fix streaming and default configuration values by @willkill07 in NVIDIA#1641
  • Add validation alias for nvext_max_sensitivity in DynamoLLM by @dnandakumar-nv in NVIDIA#1657
  • Treat explicit null defaults as nullable in MCP schema translation by @AnuradhaKaruppiah in NVIDIA#1665
  • Fix Unicode escape sequences showing in console workflow output by @yczhang-nv in NVIDIA#1664
  • Improve ReAct tool input parsing for Python-style Action Input literals by @AnuradhaKaruppiah in NVIDIA#1666
  • Add missing dependency for nvidia-nat-opentelemetry to nvidia-nat-langchain by @dagardner-nv in NVIDIA#1670
  • Fix ReAct agent parsing failures with reasoning models (<think> tags) by @yczhang-nv in NVIDIA#1667
  • Simplify the example questions to bypass priv levels by @AnuradhaKaruppiah in NVIDIA#1672
  • Fix thought matching issue in ReAct agent with the Llama-3.1-Nemotron-Nano-4B-v1.1 model by @dagardner-nv in NVIDIA#1675
  • chore(deps): provide upper-bound for starlette; bump grpcio versions by @willkill07 in NVIDIA#1669
  • Fix setting the openai base url for llama index by @dagardner-nv in NVIDIA#1686
  • examples(mcp): make example more robust to LLM hallucination by @willkill07 in NVIDIA#1695
  • Fix mixture of agent example from reaching GRAPH_RECURSION_LIMIT by @dagardner-nv in NVIDIA#1697
  • fix: handle GraphRecursionError gracefully in tool_calling_agent by @mnajafian-nv in NVIDIA#1705
  • Update nat-ui submodule to include conversation state fix by @ericevans-nv in NVIDIA#1704
  • Fix alert_triage_agent empty reports in offline mode (#1699) by @mnajafian-nv in NVIDIA#1703
  • fix(notebooks): add missing nat workflow reinstall before nat run by @mnajafian-nv in NVIDIA#1713
  • fix(simple-web-query): harden tool description; disable thinking by @willkill07 in NVIDIA#1722
  • Fix pydantic model validation for nvext hints by @dnandakumar-nv in NVIDIA#1723
  • Pin nvidia-nat-ragaai to setuptools v81 by @dagardner-nv in NVIDIA#1730
  • Fix Strands integration tests by @dagardner-nv in NVIDIA#1731
  • fix: replace llama-3.1-405b model in email phishing analyzer by @mnajafian-nv in NVIDIA#1712
  • Fix alert triage agent: switch to nemotron-3-nano model and improve prompts by @hsin-c in NVIDIA#1750
  • fix(ci): sanitize sbom license response by @willkill07 in NVIDIA#1763
  • fix: replace deprecated mistral-nemo-12b and fix reasoning agent tool discovery by @mnajafian-nv in NVIDIA#1781
  • fix: toplevel pyproject.toml should specify tool.uv.managed=true by @willkill07 in NVIDIA#1783

📝 Documentation Updates

  • Fix typo in documentation for uv sync command by @thepatrickchin in NVIDIA#1542
  • docs: 1.5 migration guide for packaging by @willkill07 in NVIDIA#1625
  • chore(docs): add GitHub Issues/PRs to Linkcheck ignorelist by @willkill07 in NVIDIA#1642
  • Ensure that we allways spell vLLM with the same casing that the proje… by @dagardner-nv in NVIDIA#1644
  • Document the need to set the NVIDIA_API_KEY in the Redis example by @dagardner-nv in NVIDIA#1678
  • Add documentation for langsmith evaluators by @pastorsj in NVIDIA#1643
  • Organize alternate source/package install commands into tabs by @dagardner-nv in NVIDIA#1737
  • Cleanup vale vocabulary by @dagardner-nv in NVIDIA#1745

Other Changes

  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1275
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1279
  • Revert langchain version update in example for dynamo by @mnajafian-nv in NVIDIA#1299
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1302
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1325
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1396
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1413
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1461
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1467
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1468
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1470
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1471
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1473
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1474
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1481
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1487
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1488
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1496
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1497
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1502
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1504
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1510
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1511
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1517
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1519
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1521
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1523
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1524
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1527
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1531
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1532
  • Forward-merge release/1.4 into develop by @rapids-bot[bot] in NVIDIA#1534

New Contributors

  • @bledden made their first contribution in NVIDIA#1564
  • @pastorsj made their first contribution in NVIDIA#1567

[1.4.1] - 2026-02-09

🐛 Bug Fixes

  • fix(serve): ensure a single event loop for python 3.11 by @willkill07 in NVIDIA#1558

📚 Documentation

  • Document new nat serve CLI flags by @dagardner-nv in NVIDIA#1562

Full Changelog: https://github.com/NVIDIA/NeMo-Agent-Toolkit/compare/v1.4.0...v1.4.1

[1.4.0] - 2026-02-02

🚀 Notable Features and Improvements

  • LangGraph Agent Automatic Wrapper: Easily onboard existing LangGraph agents to NeMo Agent Toolkit. Use the automatic wrapper to access NeMo Agent Toolkit advanced features with very little modification of LangGraph agents.
  • Automatic Reinforcement Learning (RL): Improve your agent quality by fine-tuning open LLMs to better understand your agent's workflows, tools, and prompts. Perform GRPO with OpenPipe ART or DPO with NeMo Customizer using NeMo Agent Toolkit built-in evaluation system as a verifier.
  • Initial NVIDIA Dynamo Integration: Accelerate end-to-end deployment of agentic workflows with initial Dynamo support. Utilize the new agent-aware router to improve worker latency by predicting future agent behavior.
  • A2A Support: Build teams of distributed agents using the A2A protocol.
  • NeMo Agent Toolkit Safety and Security: Strengthen safety and security of workflows by simulating scenario-based attacks, profiling risk, running guardrail-ready evaluations, and applying defenses with red teaming. Validate defenses, profile risk, monitor behavior, and harden agents across any framework.
  • Amazon Bedrock AgentCore and Strands Agents Support: Build agents using Strands Agents framework and deploy them securely on Amazon Bedrock AgentCore runtime.
  • Microsoft AutoGen Support: Build agents using the Microsoft AutoGen framework.
  • Per-User Functions: Use per-user functions for deferred instantiation, enabling per-user stateful functions, per-user resources, and other features.

🚨 Breaking Changes

  • Update weave trace identifiers by @dagardner-nv in NVIDIA#1055
  • feat: switch calculator functions to a single function group by @willkill07 in NVIDIA#954
  • Use Pydantic SecretStr fields for all sensitive values by @dagardner-nv in NVIDIA#1123
  • Migrate Zep Cloud integration from v2 to v3 API by @jackaldenryan in NVIDIA#1122
  • feat!(llm): exclude unset fields in model dump for all LLMs and Embedders by @willkill07 in NVIDIA#1143
  • Documentation Restructure by @dagardner-nv in NVIDIA#1231
  • Implement Per-User Function Instantiation by @yczhang-nv in NVIDIA#1206
  • Remove default_user_id from GeneralConfig to prevent unsafe per-user workflow sharing by @yczhang-nv in NVIDIA#1282
  • chore: update dependency package versions for 1.4 by @willkill07 in NVIDIA#1316
  • improvement: change Function Group separator to __ by @willkill07 in NVIDIA#1328
  • Refactor MCP Frontend: Move to nvidia-nat-mcp package by @AnuradhaKaruppiah in NVIDIA#1332
  • chore: update nvidia-nat-all and add documentation by @willkill07 in NVIDIA#1340

✨ New Features

  • Add DBNL Telemetry Exporter by @dbnl-renaud in NVIDIA#1111
  • Add default Phoenix session tracking support by @ericevans-nv in NVIDIA#1132
  • Add support for workflow configuration inheritance by @thepatrickchin in NVIDIA#1152
  • Add Middleware and native support for FunctionMiddleware for all functions by @dnandakumar-nv in NVIDIA#1130
  • Add support for a customizable MCP service account auth provider by @AnuradhaKaruppiah in NVIDIA#1176
  • Introduce vanna text2sql by @jiaxiangr in NVIDIA#974
  • Strands integration by @ronjer30 in NVIDIA#1063
  • NAT A2A Client & Server Support by @AnuradhaKaruppiah in NVIDIA#1147
  • Introduce Finetuning Harness for In-Situ Reinforcement Learning of Agentic Workflows by @dnandakumar-nv in NVIDIA#1221
  • Add Support for NeMo Customizer to Finetuning Harness by @dnandakumar-nv in NVIDIA#1241
  • Register per-user ReAct agent by @yczhang-nv in NVIDIA#1274
  • dynamo llm integration with examples, analysis, and custom predictive routers by @bbednarski9 in NVIDIA#1242
  • Add a bridge between NAT and A2A auth mechanisms by @AnuradhaKaruppiah in NVIDIA#1232
  • Migrate the a2a client implementation to per-user mode by @AnuradhaKaruppiah in NVIDIA#1281
  • Add weave feedback integration for chat interactions by @thepatrickchin in NVIDIA#781
  • Extend Middleware interface with pre/post invoke hooks and add DynamicFunctionMiddleware by @ericevans-nv in NVIDIA#1216
  • Agent Safety And Security Engine by @ericevans-nv in NVIDIA#1262
  • Microsoft Autogen Framework Integration [Synopsys] by @bbednarski9 in NVIDIA#1330
  • Implement per-user resource usage monitoring endpoint by @yczhang-nv in NVIDIA#1280
  • Add automatic wrappers for LangGraph Agents by @mdemoret-nv in NVIDIA#1322
  • Make All CLI Commands Plugin-Discoverable by @AnuradhaKaruppiah in NVIDIA#1346
  • feat: Add AutoMemoryWrapper agent for automatic memory management by @jackaldenryan in NVIDIA#1137
  • Add health endpoint to FastAPI server by @antoniomtz in NVIDIA#1466

🔧 Improvements

  • Add a configurable memory profiler for the MCP frontend by @AnuradhaKaruppiah in NVIDIA#961
  • Optimize retry logic with memory management improvements by @dnandakumar-nv in NVIDIA#1014
  • Refactor to make model_name an optimizable field across LLMs by @dnandakumar-nv in NVIDIA#1020
  • Added new agent and example utilizing the OpenAI Responses API by @dnandakumar-nv in NVIDIA#414
  • Include input and output messages in weave observability traces by @thepatrickchin in NVIDIA#1050
  • Allow attaching arbitrary attributes to Weave traces by @dagardner-nv in NVIDIA#1057
  • feat: nat optimizer support for Optuna GridSearch by @bbednarski9 in NVIDIA#1076
  • Lint fixes by @dagardner-nv in NVIDIA#1097
  • Make the run_workflow method a part of the core API by @dagardner-nv in NVIDIA#1098
  • Support Redis password authentication by @dagardner-nv in NVIDIA#1110
  • Update example notebook to use the run_workflow function by @dagardner-nv in NVIDIA#1113
  • Add E2E tests for Simple RAG Example by @dagardner-nv in NVIDIA#1114
  • Add E2E test for ADK demo example by @dagardner-nv in NVIDIA#1115
  • Cleanup E2E tests by @dagardner-nv in NVIDIA#1116
  • Update password fields to use Pydantic SecretStr type by @dagardner-nv in NVIDIA#1118
  • Update fastapi version by @AnuradhaKaruppiah in NVIDIA#1117
  • Support custom MCP server implementations by @AnuradhaKaruppiah in NVIDIA#1087
  • Add reference to NAT job_id in Weave evaluation attributes by @thepatrickchin in NVIDIA#1140
  • Add evaluator reasoning to Weave score logs by @thepatrickchin in NVIDIA#1141
  • Add E2E tests for notebook examples by @dagardner-nv in NVIDIA#1128
  • Add E2E test for simple auth example by @dagardner-nv in NVIDIA#1148
  • Support Unix shell-style wildcards in dataset filter configuration by @thepatrickchin in NVIDIA#1146
  • Add optional TTL configuration for Redis object store by @thepatrickchin in NVIDIA#1157
  • Local sandbox improvements by @dagardner-nv in NVIDIA#1162
  • Forward merge 'release/1.3' into develop by @willkill07 in NVIDIA#1135
  • feat: relax temperature bounds to be model-specific by @willkill07 in NVIDIA#1172
  • Update the test_lifetime_task_timeout test to not take 60s by @dagardner-nv in NVIDIA#1171
  • Ensure that the compatibility loader is removed after each test by @dagardner-nv in NVIDIA#1177
  • Add an E2E test for Simple Calculator Galileo observability example by @dagardner-nv in NVIDIA#1095
  • Improve haystack_deep_research_agent example by @mpangrazzi in NVIDIA#1170
  • Add a simple evaluate_item endpoint by @AnuradhaKaruppiah in NVIDIA#1138
  • Remove work-around for qdrant/qdrant-client#983 by @dagardner-nv in NVIDIA#1186
  • Provide a method for adding routes at the root level of the NAT-MCP server by @AnuradhaKaruppiah in NVIDIA#1187
  • Silence warnings being emitted during tests by @dagardner-nv in NVIDIA#1189
  • Work-around slow import issue for google-adk by @dagardner-nv in NVIDIA#1192
  • Remove pytest-pretty by @dagardner-nv in NVIDIA#1193
  • Add E2E test for RagaAI Catalyst by @dagardner-nv in NVIDIA#1194
  • Create TTC Functions for Multi-LLM Generation by @dnandakumar-nv in NVIDIA#1203
  • Add a Kaggle MCP usage example by @AnuradhaKaruppiah in NVIDIA#1209
  • Security and Lint updates for AgentCore Deploy by @BuildOnCloud in NVIDIA#1220
  • Add a tabular output for evaluation results by @AnuradhaKaruppiah in NVIDIA#1239
  • Update finetuning docs and add harness to workflows guide by @dnandakumar-nv in NVIDIA#1245
  • Update README for RL Example by @dnandakumar-nv in NVIDIA#1252
  • Mark wheels with a beta tag as ready by @dagardner-nv in NVIDIA#1284
  • fix: uv.lock update for nat_react_benchmark_agent by @bbednarski9 in NVIDIA#1285
  • Add rules to try and catch a bug where default='' is used for a SecretStr field by @dagardner-nv in NVIDIA#1304
  • dynamo unit test patch and cleanup by @bbednarski9 in NVIDIA#1303
  • fix: AWS AgentCore IAM policy rules and example prerequisites by @bbednarski9 in NVIDIA#1315
  • Update copyright year by @dagardner-nv in NVIDIA#1317
  • Fix: add parent-child lineage to trace/span exporter attributes by @bbednarski9 in NVIDIA#1320
  • changed to simplified system prompt and properly handle no inputs by @mnajafian-nv in NVIDIA#1307
  • Add configurable description for sequential executor by @thepatrickchin in NVIDIA#1293
  • Add early exit mechanisms for Sequential Executor by @thepatrickchin in NVIDIA#1289
  • chore: bump github actions version to v6 by @willkill07 in NVIDIA#1334
  • Implement Non-session-aware Per-user MCPClient by @yczhang-nv in NVIDIA#1253
  • Add Configuration Preservation to Evaluation Output by @AnuradhaKaruppiah in NVIDIA#1336
  • chore: bump langchain deps; regenerate uv.lock by @willkill07 in NVIDIA#1333
  • Remove stray file unintentionally added to the repository by @dagardner-nv in NVIDIA#1339
  • Rename Sequential Executor input parameter for compatibility with generate endpoints by @thepatrickchin in NVIDIA#1291
  • Implement CLI Plugin Discovery System by @AnuradhaKaruppiah in NVIDIA#1341
  • Improve Safety and Security Engine README by @ericevans-nv in NVIDIA#1344
  • Add documentation specific rules to .coderabbit.yaml by @dagardner-nv in NVIDIA#1348
  • Improves finetuning end status logging by @dnandakumar-nv in NVIDIA#1350
  • chore: update NAT UI submodule by @willkill07 in NVIDIA#1352
  • Update the build_wheel CI stage to always build wheels with matching version dependencies by @dagardner-nv in NVIDIA#1343
  • Update langsmith.xlsx to match data in langsmith.csv by @dagardner-nv in NVIDIA#1360
  • Clean up SWE-bench example: Remove unmaintained predictor and migrate to remote datasets by @AnuradhaKaruppiah in NVIDIA#1361
  • chore: update ui submodule, semantic-kernel, and langchain versions by @willkill07 in NVIDIA#1369
  • Reorganize A2A Examples for Clarity by @AnuradhaKaruppiah in NVIDIA#1368
  • add support for langchain agents that are wrapped as async context managers by @gfreeman-nvidia in NVIDIA#1371
  • chore: bump urllib3+langchain; specify werkzeug as transitive dep by @willkill07 in NVIDIA#1375
  • chore: speed up tests by @willkill07 in NVIDIA#1378
  • Decouple HuggingFace LLM provider from LangChain dependency by @ericevans-nv in NVIDIA#1367
  • Add code owners for example data directories by @dagardner-nv in NVIDIA#1379
  • Fix Windows path parsing in find_package_root by @ericevans-nv in NVIDIA#1380
  • chore: update nvidia_nat_weave > weave > fickling dependency by @willkill07 in NVIDIA#1386
  • Add a pre-commit script to ensure output cells of notebooks are cleared by @dagardner-nv in NVIDIA#1381
  • Increase the time limit for the test stage by @dagardner-nv in NVIDIA#1400
  • Expose Dask memory_limit config by @dagardner-nv in NVIDIA#1401
  • Standardize RAG service response schema parsing by @ericevans-nv in NVIDIA#1406
  • Fix/simplify event loop test by @mnajafian-nv in NVIDIA#1405
  • fix: correct ReWOO planner prompt JSON example format by @mnajafian-nv in NVIDIA#1410
  • Add pytest-timeout and set a global 5min timeout by @dagardner-nv in NVIDIA#1411
  • Expose Dask threads per worker by @dagardner-nv in NVIDIA#1408
  • Handle consecutive status check failures with retry logic in DPO trainer adapter by @dnandakumar-nv in NVIDIA#1415
  • Fix multi_frameworks workflow CI failure by @mnajafian-nv in NVIDIA#1418
  • Update middleware to use FunctionGroup.SEPARATOR for function matching by @ericevans-nv in NVIDIA#1448
  • Update A2A docs by @AnuradhaKaruppiah in NVIDIA#1447
  • Update the build_wheel CI script to test that built wheels are installable by @dagardner-nv in NVIDIA#1457
  • Add websocket MCP auth check script (no UI) by @AnuradhaKaruppiah in NVIDIA#1465
  • docs: Restore Llama config docs in simple_web_query_eval README by @mnajafian-nv in NVIDIA#1472
  • fix: dynamo multi-worker deployment shell script update by @bbednarski9 in NVIDIA#1479
  • update package versions in uv.lock; update UI submodule by @willkill07 in NVIDIA#1469
  • Update Dask by @dagardner-nv in NVIDIA#1478
  • Improve Safety and Security retail agent docs by @ericevans-nv in NVIDIA#1492
  • improvement(adk-example): update example to prefer NVIDIA NIM by @willkill07 in NVIDIA#1495
  • add name attribute to FunctionBaseConfig for workflow naming in span exporter by @bbednarski9 in NVIDIA#1482
  • chore: prefer non-required packages are manually installed by @willkill07 in NVIDIA#1516
  • chore: remove huggingface extra by @willkill07 in NVIDIA#1525

🐛 Bug Fixes

📝 Documentation Updates

🙌 New Contributors

[1.3.1] - 2025-11-07

📦 Overview

This is a minor release with documentation updates, bug fixes, and non-breaking improvements.

✨ New Features

🐛 Bug Fixes

📝 Documentation Updates

🙌 New Contributors

[1.3.0] - 2025-10-24

🚀 Notable Features and Improvements

  • ADK Support: Supports Google Agent Development Kit (ADK). Adds tool calling, core observability, and LLM integration in this release.
  • Control-Flow Agents: Sequential Executor (Linear Agent) and Router Agent now control flow patterns of tool calls and sub-agents.
  • Function Groups: Packages multiple related functions together so they share configuration, context, and resources.
  • Hyperparameter Agent Optimizer: Automates hyperparameter tuning and prompt engineering for workflows.
  • Introductory Notebook Improvements: Reorganizes getting started notebooks and adds Open in Colab links.
  • LLM Improvements
    • Adds LiteLLM Provider
    • Supports GPT-5 (/chat/completions endpoint only)
    • Adds Nemotron thinking configuration
  • MCP Improvements
    • Supports streamable-http - sse is no longer the default transport type.
    • Supports initial authorization - Enables connecting to MCP servers that require authentication.
    • Supports multiple MCP tools from a single configuration - Pulls in entire tool sets published by MCP servers or filters them based on user configuration.
    • Enhances CLI utilities for MCP servers and clients - Improves the nat mcp sub command for querying, calling, and listing tools.
  • Python 3.13 support

🚨 Breaking Changes

✨ New Features

🔧 Improvements

🐛 Bug Fixes

📝 Documentation Updates

🙌 New Contributors

[1.2.1] - 2025-08-20

📦 Overview

This is a documentation only release, there are no code changes in this release.

📜 Full Change Log

[1.2.0] - 2025-08-20

📦 Overview

The NeMo Agent Toolkit, formerly known as Agent Intelligence (AIQ) toolkit, has been renamed to align with the NVIDIA NeMo family of products. This release brings significant new capabilities and improvements across authentication, resource management, observability, and developer experience. The toolkit continues to offer backwards compatibility, making the transition seamless for existing users.

While NeMo Agent Toolkit is designed to be compatible with the previous version, users are encouraged to update their code to follow the latest conventions and best practices. Migration instructions are provided in the migration guide.

🚨 Breaking Changes

🚀 Notable Features and Improvements

  • Authentication for Tool Calling: Implement robust authentication mechanisms that enable secure and configurable access management for tool invocation within agent workflows.
  • Test Time Compute: Dynamically reallocate compute resources after model training, allowing agents to optimize reasoning, factual accuracy, and system robustness without retraining the base model.
  • Sizing Calculator: Estimate GPU cluster requirements to support your target number of users and desired response times, simplifying deployment planning and scaling.
  • Object Store Integration: Connect and manage data through supported object stores, improving agent extensibility and enabling advanced data workflows.
  • Enhanced Cursor Rules: Build new workflows or extend existing ones by leveraging cursor rules, making agent development faster and more flexible.
  • Interactive Notebooks: Access a suite of onboarding and example notebooks to accelerate agent workflow development, testing, and experimentation.
  • Observability Refactor: Onboard new observability and monitoring platforms more easily, and take advantage of improved plug-in architecture for workflow inspection and analysis.
  • Examples Reorganization: Organize examples by functionality, making it easier to find and use the examples.

📜 Full Change Log

🙌 New Contributors

[1.1.0] - 2025-05-16

Key Features

  • Full MCP (Model Context Protocol) support
  • Weave tracing
  • Agno integration
  • ReWOO Agent
  • Alert Triage Agent Example

What's Changed

New Contributors

[1.1.0a1] - 2025-04-05

Added

  • Added CORS configuration for the FastAPI server
  • Added support for customizing evaluation outputs and uploading results to remote storage

Fixed

  • Fixed aiq serve when running the simple_rag workflow example
  • Added missing platformdirs dependency to aiqtoolkit package

[1.0.0] - 2024-12-04

Added

  • First release.