Does Loops Remember Between Runs?
Does Loops Remember Between Runs? No. A workflow in Loops holds no memory of its own between runs. When a contact enters a workflow, the workflow reads that contact exactly as the record stands at that moment: the properties currently stored, the mailing lists the contact currently belongs to, and the events currently attached to the profile. Anything an earlier run worked out survives only if that run wrote it back first, as a property update, a list change, or an event the next step can check
Does Loops Remember Between Runs?
No. A workflow in Loops holds no memory of its own between runs. When a contact enters a workflow, the workflow reads that contact exactly as the record stands at that moment: the properties currently stored, the mailing lists the contact currently belongs to, and the events currently attached to the profile. Anything an earlier run worked out survives only if that run wrote it back first, as a property update, a list change, or an event the next step can check.
That distinction is easy to miss because Loops feels continuous. Contacts, lists, and workflows are all still there tomorrow, and a contact halfway through a sequence resumes at the right step. The continuity lives in records and definitions, not in a memory one run hands to the next.
What does a Loops contact actually hold?
A contact is a set of properties. An email address is the only required field, and Loops adds default properties around it: first and last name, source, user group, user ID, subscription status, mailing list membership, and opt-in status. An account can add custom properties in camelCase for anything else: a plan name, a trial start date, a product role, a usage score. Custom properties accept string, number, boolean, and date values, so a contact can hold useful structured detail, not just an address.
Two properties of that record decide what memory means here. First, a contact shows the latest value. When a property is updated, the new value replaces the old one, so the profile answers what is believed now, not what was believed last month or why it changed. Second, the contact is the trigger surface for automation. A Contact updated trigger starts a workflow when a property changes from one value to another, and it can be narrowed to fire only when the previous value matches a set value. That makes properties powerful, but it also shows their limit. A property can signal that something changed. It cannot explain what was checked before the change, which alternatives were rejected, or what the next run should do differently.
Mailing lists and audience filters work the same way. A list records membership. A filter records who currently matches a set of conditions on properties and events. Both are read fresh when a workflow runs, so they describe the audience as it stands today, not the reasoning that put anyone there.
What does the workflow itself keep?
A workflow keeps its definition. Loops workflows, called loops before the rename in May 2026, are built from six node types: a trigger, emails, timers, audience filters, branches, and experiments. There are four triggers: Contact added, Contact updated, Contact added to list, and Event received, where an event such as signUp, activated, or canceled is sent in by an app or an integration. Each workflow also holds a setting for whether a person can go through once or every time the trigger happens, and contacts leave when they reach the end or when they stop matching a filter that covers the rest of the workflow, such as upgrading before a reminder email.
What a workflow does not keep is a run journal. A branch reads properties and past interaction at the moment it runs and sends the contact down one path or the other. That decision is not stored as an explanation. An event can update a contact and start a published workflow, and that write is the one bridge from one run's work to a later run's starting point. Reading current state at the right moment is the design. Carrying a narrative forward is not part of it.
Events underline the same point. An event carries its own properties at the moment it is sent, and those values can start a workflow or update a contact. Once the workflow has acted, the durable trace is whatever now sits on the contact record. The sequence of events, checks, and near misses that led to a decision does not stay together as a story a later run can reopen.
Where scheduled AI agents hit the same wall
Add a scheduled AI agent beside a Loops account and the gap widens. The agent runs on a schedule, reviews contacts, checks who activated or stalled, drafts a follow-up, or decides which segment needs attention. Inside Loops, its work survives only where it updated a property, changed a list, or sent an event. Outside Loops, in the chat or tool where the agent actually reasoned, the full context of that run, what it checked, what it ruled out, and why it chose one action over another, has no home on the contact record at all.
The next run, whether it is the same agent tomorrow or a different tool next week, meets current values and a list of events. It does not meet the reasoning. It re-reads the same contacts, re-checks the same events, and re-derives the same conclusions, because the one place that held the full story was the previous run itself, and that run ended.
Give the runs one shared memory
Vilix AI is built for exactly that missing layer. It is cloud-hosted with zero infrastructure to run, so there is no database to provision and no server to maintain before an agent can remember. Every connected tool reads and writes the same memory over MCP, so an agent working beside Loops, a coding assistant, and a chat tool all share one state instead of separate run histories.
Vilix AI stores full conversation history, not just extracted facts. A later run can revisit what was actually discussed, checked, and decided, in order, rather than guessing from a property that only holds a latest value. That is the difference between a contact that says a user is on a plan and a memory that says why that plan was chosen, what was already sent, and what should happen next.
The starting point stays simple: Get Started for Free. Free forever, no credit card. There is a free plan forever, and a 7-day Pro trial with no credit card required. Data stays portable throughout: export everything or delete it anytime. Connect a first agent, let it save its runs, and watch the next run start informed instead of blank, before any payment decision is due.
The short answer
Does Loops remember between runs? The contact remembers, in current values. Lists and filters remember who matches right now. The workflow remembers its definition and each contact's position inside it. The run itself remembers nothing. Anything that matters for tomorrow has to be written back today, as a property, a list change, or an event, or saved in a memory layer outside the workflow. Give agents that shared memory, and each run starts where the last one finished.