Is Outloop an agency service or software?

Outloop is software for operating supported AI tools across client Workspaces. You bring the Mac, AI tools and approved provider accounts. Guided implementation is an optional, separately scoped engagement.

Can one team use AI across many clients?

Yes, using separate client Workspaces and supported configurations. Reuse the tools and reviewed client-neutral methods, not another client’s Context, accounts or private data.

Does Outloop create a shared worker memory across clients?

No. Client Context and data belong to their own Workspace. Do not interpret shared runtime choice as shared client memory or automatic cross-client learning.

How do we add a client?

Create the client Workspace, add approved Context, configure the supported systems and resource boundaries the work needs, then verify a small task and the relevant access limits.

Can we start without a sales call?

Yes. Start the self-serve 14-day trial, install Outloop on your supported Mac and follow the first-Workspace guide. Request a Workflow Review if you want help scoping a real client job.

For AI agencies and multi-client implementation firms

Run AI workers for multiple clients without rebuilding every workflow.

Reuse the methods your agency has tested. Give each client a dedicated Workspace with its own context, approved accounts, permissions and activity history.

Start 14-day trial Get implementation help

Evaluate one real workflow on your company-controlled Mac.

Setup guides + product support included · API keys stay local · Cancel anytime

outloop client access vault stays locked

Outloop client workspace and approved access

One access setup Workspace approved Runtime allowed secret_exposed:false

Last updated: November 5, 2026

In short

Outloop is Workspace infrastructure for deploying and operating AI workers.

Give each client a separate Workspace and approved access while your team reuses supported AI tools and reviewed, client-neutral methods. A shared tool is not a shared client context. Outloop currently runs on a company-controlled Mac.

Evaluate the same job for a second client

Run one approved taskOn your company-controlled Mac, configure Client A’s Workspace with its approved ad account. Ask the worker for a read-only report. Check the account and result in the execution record.
Reuse the reporting methodCreate a separate Workspace for Client B. Bring over the reviewed report instructions and checks. Supply B’s own context, files, approvals and account identifiers.
Test the account boundaryRun B’s report through the configured Google Ads API route. Verify an allowed read for B, then test a read aimed at A. Inspect the refusal before relying on the setup. This check covers the routed API request only.

Inspect recorded allow and refusal decisions or follow the first-Workspace setup guide.

Separate the client setup before work begins

Client A workspace

Client A Context and approved assets

Approved accounts and resource bindings for A

Review rules and redacted audit for A

Client B workspace

Client B Context and approved assets

Approved accounts and resource bindings for B

Review rules and redacted audit for B

Explanatory diagram, not recorded evidence. Outloop checks routed requests against the configured workspace policy; provider-specific resource bindings must be configured correctly.

Context describes what the worker should know. It does not grant system access. See how reviewed Context is reused.

Client AWorkspace

Context: Client A’s brief & rules
Accounts: Google Ads A · Meta A
Resources: CRM A · Drive A
Browser: Approved identity A
Access: Approved for Client A

Work stays with this client.

Client BWorkspace

Context: Client B’s brief & rules
Accounts: Google Ads B · Meta B
Resources: CRM B · Drive B
Browser: Approved identity B
Access: Approved for Client B

Work stays with this client.

Illustrative client setup. Each service and resource needs its own approved access; Local Chrome uses the Workspace’s configured profile; Managed Browser uses a shared profile with approved identity and access. Diagram, not recorded execution.

Your agency has a client roster. Your AI setup should reflect it.

Agencies, automation companies, GTM operators and other service businesses work inside external customer environments. Client A’s reporting task may need one ad account, CRM and Drive. Client B needs a different set, with different rules and reviewers.

Giving a tool access to the agency’s systems is only the beginning. Each task needs to know whose work it is and which resources are approved. A separate client Workspace makes that operating unit explicit.

What does Outloop handle, and what does our engineer own?

With a custom build, your team maintains the link between each worker request, the client’s approved resources and its activity record. Outloop provides that check and record for supported routes. Your engineer chooses the workflow, configures each client’s access, tests the limits and decides which actions need human review. See the implementer workflow.

Ask for evidence of the boundary.

A dedicated computer and a separate chat do not prove account isolation. With shared agency access, verify how the tool handles a request aimed at another client’s resource.

Understand wrong-client prevention and shared-account resource control.

Start with one deliverable you can inspect.

Choose a bounded task with approved inputs and a visible output: a campaign draft, a client report or a staged website update. Confirm accounts and review points before expanding repeated work.

For a performance team, use the marketing-agency workflow. For client systems delivery, see GTM engineering. For a documented site example, see Webflow work.

Models and runtimes may change while the client Workspace remains the organizing home. Supported setup is still required; full provider or machine migration is not promised. Review current Product requirements.

One agency job. From brief to a real build.

Watch an AI worker use client Context and approved assets, incorporate human review, and build campaigns inside Google Ads and Meta.

Approved client Context and brand assets
Human review before execution
Real campaign objects, kept PAUSED

Start 14-day trial Download for Mac

Create your trial, then install on your company-controlled Mac. Set up your first workspace.

67-second Ellātu preview · Google Ads + Meta · No form required

Demo brief and assumptions. Real Google Ads and Meta execution, configured PAUSED. Review emails are labeled staged reconstructions. No delivery or performance claim.

Open video directly Read the English transcript Explore the complete 8:53 workflow →

What the recorded evidence shows

The preview above follows a real campaign build in demo accounts, configured PAUSED. The allow/deny screenshots below show particular access decisions; neither is a guarantee against every possible mistake.

Approved requests and refused requests

Recorded requests in a demo workspace: approved accounts execute; requests for another client’s account or campaign are refused.

Outloop Live Activity for demo workspace 018: Google Ads and Meta Ads requests, each marked Allow with result OK, including grouped batches of 10 and 14 requests. — Demo Workspace 018 · Google Ads and Meta Ads requests
Every row: **Allow · OK** — including grouped batches of 10 and 14 requests
The approved requests returned results in this client Workspace

Outloop Live Activity for demo workspace 018: allowed google_ads and meta_ads requests interleaved with three refused requests — two marked Deny with RESOURCE_ID_NOT_ALLOWED (one google_ads, one meta_ads) and one meta_ads request marked Deny with META_OBJECT_NOT_APPROVED. — **Deny · `RESOURCE_ID_NOT_ALLOWED`** (google_ads and meta_ads) — the request named an account outside this Workspace’s approved resources; the requested action was refused
**Deny · `META_OBJECT_NOT_APPROVED`** (meta_ads) — a campaign, ad set or ad that does not belong to the approved ad account; the requested action is not sent to Meta
Approved requests in the same workspace keep executing around the refusals — **Allow · OK**

Real execution in an Outloop demo workspace (workspace 018), published in the September 26, 2026 proof pack — not a customer case study, and not evidence of ad spend, delivery or performance. Enforcement covers requests routed through Outloop. See the runtime access walkthrough for the recorded flow.

What Outloop enforces — and where the boundary is

What Outloop enforces

Approved requests are scoped to the configured client accounts and resources.
Recorded Google Ads and Meta examples show both allowed and refused requests.

What it does not guarantee

Controls apply to supported work through Outloop and depend on correct configuration.
Business decisions and publication still need your review rules.

What you still configure

Configure each client Workspace with its own Context, accounts, resources and review requirements.
Validate the workflow and result before expanding.

Start with one client workflow

You supply a company-controlled Mac, your agent and model, the provider accounts, and a reviewer. Configure one workspace and verify one bounded task before expanding.

Start 14-day trial Set up your first workspace

Still selecting the job? Get the free agency operating playbook. For implementation support, scope one workflow with a specialist.