DEVELOPMENT STATUS
All account names, campaign data, metrics, and actions shown here are fictional demonstration data. Buttons only simulate the planned workflow and never call the Google Ads API.
Application review overview
Internal advertising operations, with controls before every API action.
A Python application in development for one individual advertiser to create eligible customer accounts, build and manage Search campaigns, analyze performance, and apply validated optimization changes to accounts the developer owns or is explicitly authorized to manage.
Project stage
Design & implementation
Interface mockups for compliance review
Planned workflow
From authorized account to logged change
- 1AuthorizeOAuth 2.0 connects only owned or explicitly authorized accounts.
- 2Create or manageBuild eligible customer accounts and Search campaign resources.
- 3AnalyzeSelected non-sensitive metadata and aggregated metrics are evaluated.
- 4ValidatePermissions, policy checks, budgets, bids, and change thresholds are enforced.
- 5Apply and logEligible changes are submitted through the API and recorded in an audit trail.
Business model
Internal use only
The developer is an individual advertiser engaged in performance and affiliate marketing and may earn commissions from qualifying conversions through approved affiliate programs.
- The software is not sold or licensed.
- No clients or third parties can access the tool.
- Only owned or explicitly authorized accounts are managed.
- The developer token is never exposed through another API.
Google Ads API use
Requested functions
CREATE
Customer accountsEligible accounts beneath the authorized manager account.
MANAGE
Search campaignsCampaigns, ad groups, ads, keywords, bids, budgets, and status.
READ
Performance reportsAggregated metrics used for internal reporting and optimization.
Interface mockup 01
Eligible customer account creation
The planned form prepares a customer-account request beneath the authenticated manager account. It does not bypass account limits, policy enforcement, billing setup, or Google review.
New customer account draft
CustomerService
Draft validated for demonstration
Required fields are present. A production request would still check API permission, manager-account eligibility, account limits, and server response before recording the result.
Pre-submit controls
Account safeguards
- RequiredAuthorized MCCThe selected manager must match the configured login customer.
- RequiredCreation eligibilityThe request stops if Google returns an account-limit or permission error.
- LockedCurrency and time zoneValues are confirmed before creation because they cannot be casually changed later.
- LoggedAPI responseRequest identifier, timestamp, outcome, and error category are stored.
Interface mockup 02
Search campaign builder
AI services may draft campaign structure and advertising content. The Python application validates the result before any eligible resource is submitted through the Google Ads API.
Campaign settings
DraftResponsive Search Ad
Generated content preview
Sponsored · example.com
Explore Demo Brand Products | Shop the Latest Collection
Discover selected products from Demo Brand. Browse details and availability online.
Headline 1Explore Demo Brand Products27 / 30 characters
Headline 2Shop the Latest Collection26 / 30 characters
DescriptionDiscover selected products from Demo Brand. Browse details and availability online.85 / 90 characters
Keywordsdemo brand products, shop demo brandPhrase and exact match draft
✓ Length limits✓ URL format✓ Budget ceiling✓ Prohibited-claim scan✓ Account permission
Planned submission sequence
Atomic checkpoints01BudgetCreate or reuse an approved campaign budget.
02CampaignCreate a paused Search campaign with targeting.
03Ad groupCreate the ad group and keyword criteria.
04AdSubmit the validated responsive Search ad.
05ReviewKeep delivery paused until setup checks complete.
Interface mockup 03
AI-assisted optimization queue
Configured AI APIs, including OpenAI, return structured recommendations. The application applies only actions that pass deterministic permission, policy, budget, bid, and change-threshold rules.
Recommendation queue
2 simulated items
Low risk
Pause one low-performing keyword
Demo keyword recorded 42 clicks and no conversions in the selected reporting window.
- Action
- Set keyword status to PAUSED
- Threshold
- ≥ 40 clicks, 0 conversions
- Budget effect
- No budget increase
Budget change
Increase a daily budget by 8%
Demo campaign met the configured conversion and cost thresholds for seven days.
- Action
- $25.00 → $27.00 per day
- Threshold
- Maximum single change: 10%
- Budget effect
- Within $30.00 account ceiling
Deterministic controls
AI cannot bypass these rules
- OAuth scope and customer authorization
- Allowed resource and operation list
- Campaign and advertising-policy checks
- Daily budget and bid ceilings
- Maximum percentage change per action
- Cooldown window after a prior change
- Structured schema validation
- Complete audit-log entry
Select “Run validation preview” to simulate a rules check. No API request will be made.
Interface mockup 04
Reporting and immutable action history
The application retrieves authorized campaign data for internal analysis and records proposed, blocked, and applied changes with timestamps and reasons.
Campaign performance
Last 30 days · Demo| Campaign | Status | Impr. | Clicks | Cost | Conv. |
|---|---|---|---|---|---|
| Demo Brand | Core | Enabled | 10,250 | 742 | $486.30 | 29 |
| Demo Brand | Category | Enabled | 6,880 | 342 | $214.56 | 8 |
| Demo Brand | Test | Paused | 1,290 | 42 | $31.08 | 0 |
Audit log
Every decision is recorded
10:42:18
Simulation completedBudget rule passed · No API call
10:36:04
Recommendation blockedChange exceeded demo threshold
10:31:55
Report refreshedAggregated demo records loaded
10:28:11
Authorization verifiedMock customer permission check
Application design
Security, data handling, and compliance boundaries
Credentials stay outside source code and are never sent to an AI provider. The application minimizes external data, validates every automated action, and maintains an internal audit trail.
Planned data flow
Least data necessary1Google Ads APIAuthorized campaign and aggregated performance data
2Python controllerFiltering, validation, permissions, limits, and logging
3Configured AI APISelected non-sensitive fields only; structured recommendation returned
Only an eligible, schema-valid recommendation may return through the controller for an API action. AI output never receives direct Google Ads credentials or unrestricted execution access.
May be sent for analysis
Selected advertising data
- Campaign type and delivery status
- Ad text and keyword text required for analysis
- Aggregated impressions, clicks, cost, and conversions
- Configured budget and bid values
- Non-sensitive targeting settings
Never sent to AI providers
Protected information
- Developer tokens and OAuth client secrets
- Refresh tokens, private keys, or passwords
- Payment and billing information
- Personally identifiable information
- Unrelated customer or account data
Credential and action controls
Defense in depth
OAuth 2.0Google-supported authorization; only approved customer access.
External secretsCredentials stored outside the public repository and application logs.
Rule engineAllowlisted operations, structured schemas, budget limits, and thresholds.
Failure safeInvalid, incomplete, unauthorized, or over-limit actions stop before mutation.
Audit trailTimestamp, resource, proposed action, validation outcome, and API response.
Policy boundaryNo attempt to bypass review, suspensions, enforcement, or account limits.
Availability
No third-party access
The application is intended solely for the repository owner. It is not a public SaaS product, client portal, external API, or resold service.
Compliance statement
Application scope and disclosures
The application is intended to operate in accordance with the Google Ads API Terms and Conditions, Google Ads API policies, applicable advertising policies, and the policies of configured AI API providers. Submitting an advertisement through the API does not bypass Google Ads review or guarantee serving. Material changes to tool use will be disclosed through the appropriate Google Ads API process.