Version 1.1.0
What's New
NVIDIA Dynamo Platform
- Add NVIDIA Dynamo platform components for optimized LLM inference on EKS (#102)
- GPU Operator, Monitoring (Prometheus + Grafana), Dynamo Platform, Dynamo vLLM Serving
- AIPerf Benchmark for concurrency sweep
- AIConfigurator for auto TP/PP recommendation and SLA-driven deployment
- Supports aggregated and disaggregated inference modes with KV cache routing
OpenClaw AI Agent Orchestrator
- Add OpenClaw AI Agent component and examples (#87-#100)
- Bridge server for orchestrating AI coding agents on Kubernetes
- Document Writer Agent example (Git-native documentation generation)
- DevOps Agent example (interactive cluster management with kubectl, helm, AWS CLI)
- Open WebUI integration via auto-registered pipe functions
- LiteLLM and Langfuse integration
- Add Claude Opus 4.6 to Bedrock models for OpenClaw agents
Open WebUI Improvements
- Auto-register pipe functions on agent install/uninstall (#99)
- Pin Helm chart v12.8.1 and add admin email/password configuration (#99)
- Fix service URL and statefulset name for API calls (#99)
Bug Fixes
- fix: EKS default experience for users pre-Dynamo push (#106)
- fix: OpenClaw gateway auth, model routing, and streaming issues
- fix: OpenClaw deployment bugs and build pipeline issues
- fix: remove KEDA dependency and clean up unnecessary README sections
- fix: correct extraEnvVars to env in Open WebUI to avoid duplicate OPENAI_API_KEY
Dependency Updates
- Bump litellm from 1.61.15 to 1.83.0 (#114)
- Bump agno from 1.7.0 to 2.3.24 (#116)
- Bump fastmcp to latest (#112)
- Bump handlebars from 4.7.8 to 4.7.9 (#108)
- Bump lodash from 4.17.23 to 4.18.1 (#118)
- Bump langchain-core (#117)
- Bump requests, pygments, aiohttp, nltk, pyjwt, orjson and others
Full Changelog: v1.0.2...v1.1.0