a2a cloud
Start building
recorded on the live product

The whole platform. Seven real takes.

No mockups, no slides. Watch an agent go from an empty terminal to a live URL, get live-edited in a cloud dev box, become a shell command, answer a real call, and serve its own web app — about ten minutes, chaptered so you can jump anywhere.

a2a cloud — full demo
Build yours — a2a initBrowse live agents
0
mockups
7
scripted real-product takes
10min
empty box → live agent
100%
versioned recording scripts
chapters

Jump to any part

The video is stitched from seven takes. Each one starts at a real login prompt in a clean Ubuntu box — pick a chapter to skip straight to it.

01
a2a cloud

The platform, in a two-minute tour: landing, live registry, docs.

02
The code

One Python class becomes a typed, discoverable agent — read in VS Code.

03
Ship it, live-edit it

a2a init to a live URL, then cloud dev with hot reload on a public box.

04
Agents as commands

A live account: a2a use turns an agent into shell commands, then a real call answers.

05
The vocabulary

use, call, chat, deploy — the four commands that make agents feel native.

06
The agent is the app

One deploy serves an API, an MCP endpoint, and a real web frontend.

07
The console

The operations dashboard: your fleet, signed proofs, usage, payouts, and compliance.

“Demo videos age. Ours come from versioned scripts that run against the shipping product — every claim has a reproducible take.”
recorded by the open demo harness · narrated cue files in the monorepo
questions

What you're actually watching

Are these recordings real?

Yes — every chapter is a scripted screen recording of the actual product running in a clean Ubuntu container: real login prompts, real CLI commands, a real browser on the live site, and real agent calls against a production account. The seven takes are assembled with title cards into the full tour, and the recording harness and narration scripts are versioned in the platform monorepo.

What does the full video cover?

Seven chapters, about ten minutes. A tour of the platform and the live agent registry. An agent's code read through in VS Code. Scaffolding and deploying one to a live URL. Editing it in a cloud dev box with hot reload. Turning agents into typed shell commands and calling one. Opening a deployed agent's own web frontend. Finally the operations console: your fleet, signed receipts, usage, payouts, compliance. Jump to any chapter.

What is 'agents as commands'?

Run a2a use <agent> and the CLI fetches the agent's card — its tools and input schemas — and caches a typed stub. From then on the agent's tools are native shell commands with flags: a2a my-first-agent ask --prompt … . For one-off calls without a stub, a2a call <agent> <tool> key=value is the schema-free escape hatch.

How do I get from watching this to my own agent?

Three commands: pip install a2a-pack, a2a init my-agent, a2a deploy. The scaffold includes the manifest, a starter agent with typed tools, and an AGENTS.md so your coding agent already knows the SDK. The quickstart walks you through it end to end.

don't trust the demo — run it

Your agent could be the next recording.

pip install a2a-pack, a2a init, a2a deploy — a managed Postgres database, an MCP server, an API, a web frontend, and Ed25519-signed receipts on Agent API, public /invoke, and standard MCP tools/call executions. From empty terminal to live agent in the time it took to watch the video.

Start the quickstartRead the guides