Home/Features/Write Actions safety gate
Platform and control

Write Actions safety gate

FeedGraph is read-only towards your ad accounts until you say otherwise — then every single change still needs approving.

Join early access
How it works

A tool with your Google Ads and Meta credentials can do real damage, so FeedGraph starts unable to write to them at all. Turning Write Actions on is a deliberate, per-workspace decision, and even then nothing goes out on its own: each change is staged with a plain-English summary of what will happen, and reaches the platform only when someone approves it. Every outbound write in the codebase is forced through one enforcement point, so there is no side door — if the plumbing is missing, the request fails rather than proceeding.

Join early access →
The problem

Handing any tool your Google Ads and Meta credentials is a real risk, and the honest question is not whether the vendor means well but what happens when something goes wrong — a bad recommendation, a bug, a misread number. Most platforms answer that with a promise.

A promise is not a control. What you want to know is what the software is structurally incapable of doing without you.

How it works

01

It starts unable to write at all

A new workspace is read-only towards Google and Meta. FeedGraph can read your performance data and recommend, and it cannot change anything, because the switch that permits writing is off.

02

Turning it on is not consent to everything

Switching Write Actions on does not let changes flow. It only makes approval possible. Every individual change still has to be staged and approved on its own.

03

Each change is staged in plain English

A staged change says what will happen, to which campaign or product, on which platform, with the current value against the new one — not an action name and a payload.

04

One enforcement point, and it fails closed

Every outbound mutating request in the codebase goes through a single function. If a change has not been approved, or the approval plumbing is missing, that function throws before any request is made. There is no code path to Google or Meta around it.

05

And it is all recorded

Approvals, executions and rollbacks are written to the activity log with who approved what. Changes that support it record their previous value so they can be reversed.

What you get

  • Read-only by default, per workspace
  • Per-change approval even after writing is enabled
  • Plain-English descriptions of exactly what a change will do before you approve it
  • A single enforced choke point for every platform write, which fails closed
  • Rollback on changes that captured a previous value, and a full audit trail either way
What it does not do
  • It governs writes to ad platforms — Google Ads, Merchant Center and Meta. Edits to your own catalogue inside FeedGraph apply when you make them; those are covered by the activity log and per-field revert instead.
  • It is not a permissions system. Who may approve is controlled separately by team permissions, and approving a sensitive change requires a second factor.
  • It cannot undo everything. Rollback needs a previous value to return to, so setting a target on a campaign that had none is not reversible in one step.

Questions

What can FeedGraph do to my ad accounts without me?

Nothing. Writing is off by default, and turning it on only makes approval possible — every individual change still waits for a person. This is enforced at one point in the code that all outbound writes pass through, and it throws rather than proceeding if a change is not approved.

Can I let it run automatically once I trust it?

No, and that is on purpose. There is no autopilot mode. The value of the gate is that it is unconditional — an exception for trusted changes would be an exception an attacker or a bug could take.

Ready to try it?

FeedGraph is in early access. Apply to join and we'll onboard you personally — no credit card required.

Join early access