Brand safety & ad suitability

Step 01

Define and Align

Review representative examples and annotate the fields that need to change. Explain the edge cases to teach your team’s policy.

Output schema

SuitabilityChoose one: Placement route under the supplied advertiser rules.
suitable
manual_review
exclude
ConcernsChoose any: Applicable content concerns; empty when none apply.
graphic_injury
illegal_activity_promotion
non_graphic_violence_reporting
ExplanationText: Content evidence and advertiser rule supporting the route.

app.sutro.sh/functions/brand-safety-ad-suitability

Collecting feedback · 1/3

Assess the supplied content for ad placement under the advertiser’s rules.

Example 1 of 3

Reporting is not promotion

Model input

Advertiser rules: Synthetic advertiser rules. Exclude content depicting graphic injury or promoting illegal activity. Send non-graphic reporting on violence to manual review. Educational prevention coverage with no depictions of violence is suitable. Judge context, not isolated keywords; list only applicable concerns.
Content: Synthetic article: Police reported a robbery downtown. The article describes the investigation without graphic imagery, instructions, or endorsement.

Model output

Suitability

ConcernsMultiple values

Explanation

Why? Optional guidance

Step 02

Model Sweep

Compare models on the same expert annotations, then choose a quality–cost tradeoff.

Compare models on expert agreement and cost. The interactive chart uses illustrative measurements.

Illustrative sweep · select Value or Best to use that model in your function.

Step 03

Deployment

Call Sutro Functions like any other software. Swap the model as soon as something more performant is released.

Selected modelWaiting for sweep results.

Real-time

import sutro as so

input_data = {
    "advertiser_rules": "Synthetic advertiser rules. Exclude content depicting graphic injury or promoting illegal activity. Send non-graphic reporting on violence to manual review. Educational prevention coverage with no depictions of violence is suitable. Judge context, not isolated keywords; list only applicable concerns.",
    "content": "Synthetic article: Police reported a robbery downtown. The article describes the investigation without graphic imagery, instructions, or endorsement."
}

result = so.run_function(
    "brand-safety-ad-suitability",
    input_data,
)

curl --fail-with-body -X POST "${SUTRO_API_URL%/}/v1/functions/brand-safety-ad-suitability/run" \
  -H "Authorization: Key $SUTRO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "input": {
    "advertiser_rules": "Synthetic advertiser rules. Exclude content depicting graphic injury or promoting illegal activity. Send non-graphic reporting on violence to manual review. Educational prevention coverage with no depictions of violence is suitable. Judge context, not isolated keywords; list only applicable concerns.",
    "content": "Synthetic article: Police reported a robbery downtown. The article describes the investigation without graphic imagery, instructions, or endorsement."
  }
}'

Batch

import sutro as so

job_id = so.batch_run_function(
    name="brand-safety-ad-suitability",
    data="brand_safety_ad_suitability.csv",
)
results = so.await_job_completion(job_id)

curl --fail-with-body -X POST "${SUTRO_API_URL%/}/v1/batch-inference" \
  -H "Authorization: Key $SUTRO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "brand-safety-ad-suitability",
  "inputs": [
    {
      "advertiser_rules": "Synthetic advertiser rules. Exclude content depicting graphic injury or promoting illegal activity. Send non-graphic reporting on violence to manual review. Educational prevention coverage with no depictions of violence is suitable. Judge context, not isolated keywords; list only applicable concerns.",
      "content": "Synthetic article: Police reported a robbery downtown. The article describes the investigation without graphic imagery, instructions, or endorsement."
    }
  ]
}'