Show HN: New Agentic AI Framework in CNCF | Hacker News Reader

Show HN: New Agentic AI Framework in CNCF

github.com

23 pointsyaronsc10 comments

Hi, I'm a Dapr CNCF project maintainer. We've recently released Dapr Agents which provides agentic AI features together with built-in durable execution to guarantee statefulness and reliable agentic workflows that run to completion and retry upon failure. It runs natively on Kubernetes, has built-in OTEL integration and uses a lightweight architecture where agents scale to zero, allowing you to run thousands of agents on commodity hardware. It'd be great if you can test it out and give us feedback.

Open articleView on HN

A breakdown of how this compares to other popular agentic AI frameworks (which appear to be multiplying by the day) would be helpful - a decision guide of "is this the right package for me", perhaps

Makes total sense and we have this as WIP

I feel like that at this point, we need an AI agent to compare AI agent frameworks.

And benchmarks. Well thought out, structured, non-cherry-picked benchmarks to highlight which framework does well in what area.

Benchmarks are WIP. We're thinking about durability, task latency, agent throughput. What else would you like to see?

Pass^k and not Pass@k (see https://www.philschmid.de/agents-pass-at-k-pass-power-k). Would be a great twofer to see the code used to run the benchmarks as examples.

Will take a look, thanks!

This space is crowded and largely undifferentiated. I'd look for MCP support, state management, observability, tests, and a high bus factor to stand out. By these measures, DAPR is not competitive; it would not make my shortlist.

Assuming you knew there are lots of alternatives, what led you to create it?

Dapr's state management API supports more than 20 different databases, with the ability to stream data at high scale right into the agent tasks. That's a major differentiator as you can plug in almost any database and message bus into Dapr Agents.

We wanted to create a vendor neutral framework that doesn't over pivot on features that are tied to the backing of a commercial product. The other and no less important point is the ecosystem that Dapr has around messaging and state integrations. A lot of the Agentic AI frameworks you see today will not withstand a restart of the process,let alone complete cluster shutdown. Dapr has durability built-in to handle these catastrophic failures

I'd lead with durability and streaming, back it with benchmarks, and suggest some use cases (e-commerce, etc).

So I might have titled this submission "A durable, streaming agent framework".

Good suggestion, thank you

Reply on news.ycombinator.com