Dapr
CNCF microservices runtime with Dapr Agents for production AI agents.
About
Dapr is the open-source distributed application runtime originally created at Microsoft and now under the CNCF. In 2025 the project shipped Dapr Agents, a Python framework built on Dapr's durable workflow engine that gives AI agents production-grade reliability: automatic retries, persistent state across dozens of databases, secure multi-agent coordination, and built-in observability.
Dapr Agents reached v1.0 general availability after collaboration with NVIDIA and the Dapr community, targeting teams that need to move AI agents from prototypes to dependable production systems on Kubernetes. It integrates with providers such as OpenAI, AWS Bedrock, and Hugging Face, and lets developers switch model providers without code changes.
What it can do
- Durable long-running agent workflows with failure recovery
- Persistent agent state across 30+ databases
- Secure multi-agent coordination with SPIFFE identity
- Provider-agnostic LLM integration
- Built-in observability and monitoring
Setup
Install the Dapr CLI and runtime, then pip install the dapr-agents package to build stateful agents in Python.
Prerequisites
Kubernetes cluster or local dev environment; Dapr runtime installed; LLM provider API access.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 132; Trending score 80/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.