Open on ProductHunt

Blocks.ai

The Control Plane and Network Layer For AI Agents

Developer ToolsArtificial IntelligenceGitHubSDK
27votes11commentsLaunched 23h agoWebsite

By Kayley Smith, Joshua Carr, Dawid Urbas, Markus Kohler

Sonnet 4.6 fasst Produkt + Kommentar-Tonalität in vier Sätzen zusammen (~0.5 ct, einmal pro Post).

About

PubNub introduces Blocks.ai today, a control plane and global network to connect and control agents across all agent frameworks, providers, and APIs. Blocks Network supports all AI agent use cases without opening inbound ports, setting up tunnels, changing DNS, or modifying firewall rules. For more than a decade, PubNub has been the infrastructure and platform for real-time connectivity supporting billions of devices, now, PubNub delivers Blocks Network connecting the Internet of Agents.

Top comments6 of 11

  • [REDACTED]·40h ago36
    Your agent is stuck. The brain is there, but you can only reach it from a terminal or one chat app. The second you want a real frontend for it (a web page, your phone, a CLI) or to connect it to other agents, you're facing port forwarding, a tunnel, DNS, a static IP. Blocks removes that ceiling. Connect your agent once, reach it from anywhere and any frontend you build — private, free, no, ports. Then plug it into a network of agents fortified with Zero Trust security and A2A protocols. Your agent opens a single outbound connection and becomes reachable to UI, agents, users through it. No inbound ports, no DNS, no firewall changes. Then you call your own agent from any frontend (or from another agent) anywhere, and it stays private and free to you. - Wrap your existing agent in a thin handler + agent-card, then publish it and run it. - Call it from the browser with a browser-safe TaskClient + a tiny token proxy (your API key never touches the client). Add `pipe` + streams for a live chat UI. - Useful without requiring other agents (though you can always connect it to others when you need to). From day one, your agent has the frontend and capabilities it never had before. - Your agent is no longer isolated. It still runs on your machine, your model, and your data. Blocks Network just routes the task down the outbound channel and streams the result back. The same connection lets your agent call other specialists (code review, data extraction, transcription, anything) and your own other hosted agents. Add it to an MCP-native runtime like OpenClaw by registering Blocks as an MCP server, or call directly with the open SDKs (Apache-2.0, TS + Python). It's not a replacement for your runtime. Your code, your model, and your stack stay exactly as they are. You're just adding a connection. Built on PubNub (99.999% SLA, SOC2, GDPR), so the streaming chat UI and streaming specialists (i.e live transcription, live monitoring) fan out one stream to many consumers in a way that's hard to do economically yourself. Reach quickstart: https://blocks.ai/docs/quickstart SDKs (Apache-2.0): https://github.com/blocksnetwork... I'd love to hear how you currently reach your agents. That's the most useful feedback I can get today. Ask me anything.
[REDACTED]
·
19h ago
13

Interesting positioning. A lot of agent infrastructure focuses on orchestration after agents are deployed, while you're solving the connectivity layer first.

The outbound-only approach removes a lot of deployment friction, especially for teams behind strict firewalls. I'm curious how you handle identity and trust as the network grows. If one agent calls another, what mechanisms verify the caller's permissions and prevent an agent from impersonating another or accessing capabilities it shouldn't?

  • [REDACTED]·20h ago11

    How does Blocks.ai actually handle authentication and agent identity when you're stitching together agents from different frameworks, especially around secure handoffs?

  • [REDACTED]·13h ago5

    How does the pricing actually work for higher message volumes, especially across multiple agent frameworks at once? Trying to figure out if it scales reasonably or gets painful fast.

  • [REDACTED]·11h ago2

    How does pricing scale when connecting agents across multiple providers and frameworks, and is there a noticeable latency hit compared to running everything inside a single VPC?

  • [REDACTED]·5h ago0

    the per-task token rotation makes sense for discrete tasks, but what about the long-lived streaming cases you mentioned above (live transcription, voice, video)? does a task token just live for the whole duration of an open stream, or does it get silently re-issued on a timer mid-stream without disrupting the connection? asking because re-auth mid-flight is usually where these systems either add a hiccup or get it invisibly right.