Introducing State Machines. The first infrastructure to spin up enterprise environments. Any enterprise app, recreated for agents. Run thousands of environments in parallel, each with its own state.

Genuinely curious, what does this have to do with actual state machines?

such a smart idea

safe and reliable infra for agents

spinning up isolated stateful API replicas for parallel evals without paying for live enterprise seats. , this is good.

the future of ai

such a huge unlock for anyone building agents

safe and reliable infra for agents to run in

Looks solid!! 🔥

just the start

Most epic launch video

if this means the bot can click around Slack and Salesforce without breaking the real one, I’m listening

its just the beginning

if you have a salesforce api you have my credit card

just the start

This launch video is insane lol

marketing guys are insane

AI agents are going to need more than just a model. They need an environment to actually do things. This feels like a step in that direction.

the next phase of infra

SolidWorks please ser

This is exactly what you need to do a full enterprise deployment

cant wait working with you guys!

LFG!!!

this is fire 🔥🔥🔥

thanks

Amazing!

enterprise apps recreated for agents... didnt know i needed this

were going hard

this is dopee!

safe and reliable

god I've tried to build this internally, how can I get access?

pls book a meeting

Let’s goooooo @IdanBier save me a slot on the channel

It’s like having a thousand hands, one for every task

just the start

An impressive infrastructure will undoubtedly be the future

safe and reliable

that's absolutly insane

thanks!

Noiceee

Need this so bad!!

lfggg

This gives teams a practical way to validate agents before putting them in real environments.

agents need a place to practice

Wow!!

thanks!

Thousands of parallel envs each with its own state is wild

the future of ai

🔥

lfg

Let’s fucking gooo

lfg ha?

Replicas of Salesforce / Zendesk / SAP for agent eval are overdue. The hard part starts after the sandbox: production writes into the real SoR with identity, policy, and an audit trail. Eval environments prove the loop. A governed gateway ships it.

Exactly. When an agent can interact with the same third-party app through the UI, API, MCP, webhooks, or browser automation, a thin mock won’t tell you whether it actually works. You need a realistic, stateful replica to prove the loop, then a governed gateway to run it safely in production.

@GuttmanIta76973 Browser tests need state too: profile, origin/account/tab, pending action and approval receipt, then fresh read-back after reconnect. Otherwise a governed gateway can authorize a stale UI write. I build Hronaut, a local browser MCP.

@statemachineshq Exactly. Browser state is part of the write contract: profile, pending action, approval receipt, then read-back. Otherwise the gateway authorizes a stale UI mutation.

Stateful environments could make agent testing dramatically more realistic

were going hard

Hey, please consider me for the collaboration too. I’d love to work with you! 🙌

amazing one guys

thanks!

This is an important layer for anyone building agents across multiple integrations.

safe and reliable infra for agents to run on

if this is stateful i need an sfdc one rn

lfgg

Realistic testing is becoming essential as agents move into enterprise workflows.

its just the start!