Terms of service

What this covers

These terms apply to the organisation that signs up, not to the individual who happens to click the button. The people who work in that organisation get access through it, and the organisation is responsible for what they do with that access.

An organisation's workspace is separate from every other organisation's. That separation is not a policy we promise to observe — it is enforced in the database itself, on every query, so a mistake in application code cannot expose one organisation's records to another.

Who can hold an account

Anyone acting for an organisation that can enter into an agreement. If you are signing up on behalf of a company, you are confirming you are allowed to.

Who can be added to one

The organisation decides. Adding someone gives them access to that organisation's data at the level their role allows, and removing them ends it.

One account is not a shared login

Credentials belong to a person. Sharing one removes the only record of who did what, which is the record both of us need if something goes wrong.

Sending, and what the product refuses to do

You are responsible for having a lawful basis to contact the people you contact. That is not boilerplate: it is the one obligation here that we cannot discharge for you, because only you know where a contact came from.

What we can do is refuse to help you get it wrong, and that is how the product is built. Consent, suppression and jurisdiction checks run before a send rather than being reported after one, and there is no configuration, plan or support request that turns them off.

An unsubscribe is permanent and immediate

Someone who opts out is suppressed at the point of sending. Re-importing them does not clear it, and neither does creating a new list — suppression is checked against the address, not against the list it arrived on.

Permission is never inferred from a relationship

Having someone's address, having met them, or having a record of them in a CRM is not recorded as agreement to be contacted.

Purchased and scraped lists are not acceptable use

Not a taste preference. A list nobody on it asked to be on generates the bounce and complaint rates that damage the sending reputation shared with every other customer.

Where you are sending changes what is allowed

Rules differ by jurisdiction, and the stricter rule is applied rather than the one most convenient for the sender.

A test never sends anything real

Test and development environments cannot reach a real recipient. Outbound sending in those environments goes to a substitute, not to a person.

The rest of acceptable use

Do not use the product to break the law, to impersonate someone, to attack or probe systems you are not authorised to test, or to work around another organisation's security. Do not resell access, and do not use it to build a competing product from the inside.

Do not upload content you have no right to, and do not use the product to process categories of personal data it was not designed for — health records, payment card numbers, government identity numbers and children's data are the ones worth naming, because a customer usually reaches for them without thinking of them as unusual.

Text the product drafts for you

The product drafts text — messages, summaries, suggested next steps. A draft is a starting point, not an approval, and anything sent under your organisation's name is your organisation's responsibility even when software wrote the first version of it.

Where a person on the receiving end would reasonably need to know that they are interacting with software rather than a colleague, the product says so. That obligation is not something a customer can switch off to make a message read better.

When Runner supplies the AI, rather than you connecting your own key, we choose which AI model and vendor handles each task, and we may change that choice at any time without notice, so the work runs on whichever does it best. The vendors we may use are named on our providers page. Before we use one that page does not name, we update it and email your account's owners first. When you connect your own key, the vendor is yours to choose.

You can still choose the model yourself

A team member can be set to a specific model. When Runner supplies the AI, that member then uses credits at that model's rate.

Nothing outbound goes without a human approving it

Approval is a step in the product, not a setting. The point of a draft is that someone reads it.

A generated answer can be wrong

Treat anything drafted as you would a first pass from a new colleague: useful, and checked before it leaves.

Your data stays yours

Everything you put into the product — your contacts, your messages, your records of what happened — remains yours. Using the product does not transfer ownership of any of it, and ending the agreement does not either.

We use it to run the service for you: to store it, to show it back to you, to send what you ask us to send, and to fix things when they break. We do not sell it.

One use reaches beyond your account. When someone identifies themselves on a site you connect, for example by submitting a form or signing in, Runner adds the identifiers involved to an identity graph it operates across its customers, and may use that graph to recognise the same person elsewhere, including for other customers. You permit this use, and you are responsible for telling the people who use your sites about it.

Another use reaches beyond your account, and you can switch it off. Runner may use what your organisation connects, such as product catalogues, code repositories and documents, and de-identified measures of how the product performs for you, to improve Runner and to produce aggregate insights that help every customer. It never uses your contacts or other personal data this way, never shows your content to another customer, and does not train AI models on it. Data you connect from another service is used this way only where that service's terms allow it, and data from Google services never is.

An owner or admin can turn this off at any time in Settings, Organization. Turning it off stops any further use; insights already produced identify no one, so there is nothing in them to withdraw. Recommendations tuned to workspaces like yours may then be less relevant.

Your data is not training data

Your content is not used to train models for anyone else's benefit.

You can take it out

Export is a normal function of the product rather than a request that has to be approved, and it is not withdrawn when an account is closing.

The software itself is ours

You get the right to use the product for as long as the agreement runs. That right does not include copying it, taking it apart, or rebuilding it.

Security, and who else is involved

One organisation's data is separated from another's in the database itself, on every read and every write, apart from the identity graph described above. Credentials you connect — a mail-sending role, a calendar, a payment provider — are stored encrypted, and the parts that are secret are never returned to a browser and never written to a log.

Running this product involves other companies: somewhere to host it, somewhere to send mail from, providers whose models draft text. They are named, individually, on the sub-processors page rather than described as a category.

You keep control of the credentials you connect

A connection you authorise can be revoked by you, and revoking it stops the access rather than merely hiding the button.

We tell you when something happens to your data

If your data is affected by a security incident, you are told — with what is known, when it is known, rather than after it has been tidied up.

Who at Runner can read your content

A few Runner staff can reach what is stored in a workspace, so they can run the service and fix it when it breaks. They do not look at it unless you ask us to for support, it is needed to investigate a security problem or abuse, or the law requires it.

Availability and support

The product is provided as it is. It is actively run and actively fixed, and problems that affect sending or data are treated as the most urgent kind, because they are.

These terms do not state an uptime percentage or a guaranteed response time. That is deliberate. A number here would be a commitment, and publishing one we do not measure and could not evidence would be worse than saying plainly that no such number is promised yet. If your organisation needs one, it belongs in a written agreement with us rather than on this page.

Maintenance happens. Where it will interrupt something, notice comes first where that is possible; where a fix is urgent enough that waiting is the greater risk, it is applied and then explained.

Limits

Neither of us is responsible to the other for indirect losses — lost profit, lost opportunity, or the cost of something that did not happen because a message did not arrive.

Nothing in these terms limits anything that cannot lawfully be limited, and that carve-out is not decoration: it is the part that survives wherever local law says a limit does not apply.

You are responsible for the consequences of what you send and who you send it to. We are responsible for running the product as described here.

Free tools

The free tools on our website, including the Menu Translator, are provided free and as they are: their results, translations included, can be wrong, you are responsible for checking a result before you use or publish it, and we are not responsible for loss caused by an error in one, except where the law does not allow that limit.

메뉴 번역기를 포함한 웹사이트의 무료 도구는 무료로, 있는 그대로 제공됩니다. 번역을 포함한 결과는 틀릴 수 있고, 결과를 사용하거나 게시하기 전에 확인할 책임은 이용자에게 있으며, 법이 허용하지 않는 경우를 제외하고 결과의 오류로 인한 손해에 대해 Runner는 책임지지 않습니다.

Fee on managed ad spend

From 1 December 2026. When you let Runner manage a product's ads inside a budget you set, we charge a fee only on managed ad spend above a monthly threshold.

The amount

No fee on the first US$2,000 of managed ad spend in a calendar month (UTC), counted across all of your organisation's products. Above that, 5% of the managed spend over US$2,000.

What counts

Spend your ad platforms report on campaigns while an active managed ad budget in Runner covers them. These do not count: campaigns you run yourself, spend while a budget is paused, used up, expired or ended, and anything you pay Runner for, such as your plan or add-ons.

Your ad money stays yours

Each ad platform bills your own ad account. Runner never holds or pays your ad spend.

Worked out after the month

A month closes for billing seven days after it ends. A day the platforms report late is added to, or taken from, the next month's fee. We never send a second bill for a closed month.

Shown before it is charged

The threshold and rate appear in Runner wherever you set or raise a managed budget, and Settings → Usage and credits shows the month's managed spend and the fee so far. The fee appears as one line on your next invoice.

If the terms change

A new threshold or rate applies only to months that start on or after its date. A month already under way keeps the terms it started with.

When these terms change

These terms carry a version number and a date, both at the top of this page. A substantive change bumps the version, and the date says when. That is deliberate: a document that can be edited silently is not a document anyone can rely on.

A change that materially reduces what you get, or materially increases what you owe, is notified before it takes effect rather than announced by having already happened. If you do not want to continue under the new version, you can stop and take your data with you — the export described below is not withdrawn as leverage.

Money-back guarantee

14-day money-back guarantee. An owner or admin of your workspace can ask for a refund of your first plan payment, monthly or yearly, within 14 days of paying, from Settings → Team. It covers a workspace's first payment for a Runner plan, made in US dollars. Later plan payments and bought credits (top-ups and auto-reloads) are not covered, each workspace can ask once, and each card can be refunded once.

We refund what you paid, less what you used that cost us money: plan credits you used, at $1 for every 100 credits, and the AI, pictures, video, voice and presenter rendering we ran on our own account for you from your payment until your refund, at the same rates as AI credits. That includes promotional credits we gave you and usage past the credits you had. Bought credits are never deducted and are not refunded; they stay yours until they expire. We cover the card processing fee. If what you used covers what you paid, there is nothing to refund.

When a refund is approved, your plan ends straight away: the subscription is cancelled with nothing more to pay, and unused plan credits end that day. The money goes back to the card you paid with. Telling us why you're leaving is optional.

Ending it

You can stop at any time. We can stop the account for non-payment, or for the sending conduct described below — sending to people who did not agree to hear from you puts the sending reputation of everyone else on the same infrastructure at risk, so it is the one thing that ends an account quickly.

Ending the agreement is not the same as deleting your data, and the two are deliberately kept apart. Your data remains exportable for a period after the account closes, because a customer who leaves in a hurry is exactly the customer most likely to need it back.

Suspension is narrower than termination

Where the problem is a specific behaviour, the narrower response is used: sending can be paused for one project without closing the account or touching anything else.

Deletion is a request you make, not a thing that happens to you

Closing an account does not immediately erase it. Ask for deletion and it is carried out; until you do, the data stays retrievable.

Which law applies

These terms are governed by the laws of the Republic of Korea. Disputes go to the Seoul Central District Court, unless the law of the place you are in gives you a court closer to home that you are entitled to use instead.

Choosing a law does not switch off anyone else's. Data protection rules follow where you and the people in your data are, not what this page says — so if the GDPR, the Korean Personal Information Protection Act, or your own state's privacy law applies to you, it still applies, and nothing here reduces it.

Where we are

The organisation behind this product is established in the Republic of Korea. That is why Korean law governs — it is where we actually are, rather than a jurisdiction chosen to be inconvenient to reach.

Rights you cannot sign away

Wherever local law says a protection cannot be waived by contract, it is not waived by this one. That carve-out is the part that survives everywhere.

Text message reminders

Runner meeting reminders is an optional text program. A booking page can offer one text before a booked meeting, and you get it only if you tick the box. You can book without it.

How many

One text for each meeting you ask to be reminded of.

What it costs

Nothing from us. Message and data rates from your carrier may apply.

Stopping and help

Reply STOP to opt out, or HELP for help. You can also email team@runtheworld.ai.

Where it works

US mobile numbers only. Carriers are not liable for delayed or undelivered messages.

Your number

How your mobile number is handled is in the privacy statement.

Asking about any of this

If something here matters to your organisation and this page does not answer it, ask before you sign up rather than after. A term you had to guess at is a term neither of us can rely on.

What happens to personal data — what is collected, how long it is kept, and what a person can ask for — is set out in the privacy statement rather than repeated here, so that the two cannot drift apart.

Who you are contracting with

Registered name

SHMAPLEX

상호

슈마플렉스

Business registration number

208-06-17257

Representative

Robert Sawyer

Address

서울특별시 강남구 선릉로 428, 2층