Research Preview

Ryu. A managed platform for AI agents.Built on open-source infrastructure.

Ryu is a managed platform for AI agents, built on open-source infrastructure. Start without building from scratch, add tools with plugins, and turn your agents into apps. Monitor every run and control what your agents can access.

Get started See how it works →

Running

support-agent

Triage inbox

Needs approval

ops-agent

Merge release PR

Completed

sales-agent

Update CRM

Every run monitored. Sensitive actions wait for you.Agents run around the clock; you stay in control.

How Ryu works

Five things Ryu handles for every agent you run, so you can focus on what the agent actually does.

Turn your agents into apps

Give your agents an interface your team and customers can use.

Start without building from scratch

Ryu is built on open-source infrastructure, so you never start from scratch.

Your agent

Plugins (12)

Memory

Open-source runtime

Add tools with plugins

Give your agents the tools they need by adding plugins.

Slack

GitHub

Postgres

Control what your agents can access

Decide which tools and data each agent can reach, and require approval where it matters.

Support agent

read

Gmail

Ops agent

approval

GitHub

Sales agent

denied

Billing

Monitor every run

See every run step by step, from trigger to tool call to result.

Run startedTriggered by Slack

PlanModel call

Tool callGitHub · read

ApprovalAccess checked

CompletedLogged

Run startedTriggered by Slack

→

PlanModel call

→

Tool callGitHub · read

→

ApprovalAccess checked

→

CompletedLogged

Get started without building from scratch

From an empty project to an agent your team relies on, in three steps.

Start from a working agent

Ryu runs on open-source infrastructure, so your agent gets a runtime, memory and hosting from day one. Bring your own agent or start from one that already works.

Open sourceRuns 24/7Cloud or self-hosted

Plug in the tools it needs

Add tools with plugins instead of writing integrations. Connect Gmail, Slack, Notion, GitHub and more, and choose exactly what each agent is allowed to do with them.

PluginsAccess rulesApprovals

Ship it as an app

Turn your agent into an app your team or customers can use, then watch every run as it happens and step in when something needs a human.

Agentic appsRun historyAlerts

One platform, every piece

Everything an agent needs to run in production, managed for you.

Ryu Cloud

Deploy agents and keep them running 24/7.

Ryu Connect

Plugins that give agents access to the tools they need.

Ryu Gateway

Access rules and approvals before agents act.

Ryu Console

One place to see every agent and every run.

Ryu Apps

Turn agents into apps: sites, inboxes, boards and docs.

Ryu Passport

A verified identity for each of your agents.

Ryu Box

Workspaces where agents build and preview their work.

Ryu Mail

An inbox your agents can send and receive from.

Ryu Notify

Events and alerts when runs finish or need you.

Who is Ryu for?

If you want agents doing real work without building the platform underneath them, Ryu is for you.

Teams

Put agents to work without running servers

Hand repetitive work to agents that run around the clock, with the tools and rules your team already uses.

Inbox triageStandup summariesCRM updates

Builders

Ship agent products, not infrastructure

Skip the plumbing. Start on open-source infrastructure, add plugins, and turn your agent into an app people can use.

Customer-facing appsInternal toolsAgent templates

Operators

Know what every agent did, and why

Monitor every run, set what each agent can access, and require approval before anything sensitive happens.

Run historyAccess rulesApprovals

Why Ryu

Most AI agents never make it past the demo. The agent itself is rarely the problem. Everything around it is: hosting, tools, permissions, and knowing what it actually did.

Ryu is that layer, managed for you, so your agents can do real work.

Without Ryu

Glue code, cron jobs, API keys everywhere

With Ryu

Plugins, managed runtime, scoped access

Most AI agents didn't deliver. We build the ones that do.

A Major

What you get

Built on open-source infrastructure

Agents run 24/7 in the cloud

Cloud, self-hosted or team setups

Plugins for the tools you already use

Approval before sensitive actions

Every run monitored and logged

A verified identity for every agent

Turn any agent into an app