Everything Qoren does for your agents

Each environment is a private computer your agents live and work on. Pick a size by how many agents it holds: 1, 2, 5 or 10.

Run agents on any of the three runtimes, from the same console, with the same vault and the same credits.

Give an agent scheduled tasks, see what runs next, and pause a task without deleting it.

Wake an agent when cal.com, GitHub, Stripe or any signed webhook fires. Pick events, write rules, test it and read the delivery log.

Agents in the same environment can message each other and hand off work, and you can read every message they exchange.

Move an agent to a bigger environment or another region, keeping its memory, files, keys, mailbox and spend history.

Every agent is backed up about every four hours when something changed. Back up on demand, download a copy, or restore an earlier point.

API keys and credential files are encrypted at rest and reach only the agents you hand them to. Every reveal of a value is logged.

Each agent's health is checked every few minutes. A repair reinstalls its runtime and keeps its memory and keys.

Removing an agent keeps a final backup, so you can bring it back later on any environment.

Download a copy of your account data as a JSON file from Settings whenever you want it.

Set a trigger to Propose only, or put a Hermes agent in approval mode, and approve or deny each action from one Approvals page.

Choose which Qoren tools each agent may call. Tools that can do lasting damage, like removing an agent, always ask first.

See what an agent is doing now, what it needs you to approve, what it did in the last day and what it will do next.

Open a live terminal inside an agent's workspace from its Shell tab when you want to look under the hood.

Upload, write, edit and download the reference files an agent works from, or upload a whole .zip at once.

Agents spend the monthly grant first, then credits you topped up, and pause at zero. There is no open-ended metered billing.

Automatic top-ups, usage alerts by email or webhook, and a monthly budget that can pause agents.

A statement of credits in and out by category for any date range, with grants, top-ups and expiries.

Run every agent on your own OpenRouter key, on any plan, so model usage is billed by OpenRouter instead of in credits.

Sign a Codex agent in to your own ChatGPT plan with a one-time device code, from the console or the CLI.

Give an agent its own email address to read, reply and send reports from. A mailbox is a flat monthly charge in credits.

Message a Hermes or OpenClaw agent through its own bot, and have its scheduled results land there too.

Every agent gets a Qoren tool server to look itself up, message other agents, share files as links and read usage, with only the tools you switch on.

Write reusable know-how once as a SKILL.md with its files, then hand it to any agent that needs it.

Ready-to-deploy agents, each with a job, a schedule and a deliverable. Search and filter the gallery, then deploy one or customize a copy.

Save a persona, runtime, size, scheduled tasks, files and tools as a reusable template of your own.

Move a Hermes, OpenClaw or Codex agent off a VPS or laptop. One script exports it, and Qoren deploys it with its memory and skills.

Script environments, agents, secrets, scheduled tasks and webhooks from the qoren command line or @qoren/sdk.

Invite teammates to work on the same environments, agents and clients. Seats are free; your plan sets how many people can join.

Group environments and agents by client, pick a client for each new agent, and keep a client's keys for that client's agents only.

See what each client cost you over a period, split into hosting, models, web search and mailboxes.

Quote a client for AI agents with your platform cost and margin in view, then print the proposal under your own brand.