TRam Image Reviewer
Pick exactly what to detect in an image — people, children, faces, license plates, screens, identifying documents, whiteboards — and the agent finds every instance, redacts it, and gives you a detection report with confidence and reasoning per finding.
1. Configure
Each detection's bounding box is inflated by this percentage before redaction. Higher = more forgiving against slightly-off boxes.
Free demo: 3 reviews per session, 10 per IP per day. Sign up for a real API key. Production traffic? See API docs.
2. Result
Pick categories, drop an image, and click Review & redact.
Reviewing the image…
Every detected instance, with its category, location, and confidence.
| Category | Description | Location | Confidence |
|---|
Per-category counts
Notes
3. Use it from your code
Same endpoint your application would call.
curl -X POST https://api.tramenterprise.com/v1/agents/image-reviewer/execute \
-H "X-API-Key: tk_live_..." \
-F 'input={
"categories": ["faces", "license-plates", "screens-displays"],
"padding": 0.15
}' \
-F "file=@photo.jpg"import requests, json
payload = {
"categories": ["faces", "license-plates", "screens-displays"],
"padding": 0.15,
}
with open("photo.jpg", "rb") as f:
response = requests.post(
"https://api.tramenterprise.com/v1/agents/image-reviewer/execute",
headers={"X-API-Key": "tk_live_..."},
files={"file": f},
data={"input": json.dumps(payload)},
)
result = response.json()
print(f"Found {result['output']['detection_count']} detections")using var http = new HttpClient();
http.DefaultRequestHeaders.Add("X-API-Key", "tk_live_...");
var input = """
{ "categories": ["faces","license-plates"], "padding": 0.15 }
""";
using var form = new MultipartFormDataContent();
form.Add(new StringContent(input), "input");
form.Add(new StreamContent(File.OpenRead("photo.jpg")), "file", "photo.jpg");
var response = await http.PostAsync(
"https://api.tramenterprise.com/v1/agents/image-reviewer/execute",
form);
var result = await response.Content.ReadFromJsonAsync<JsonElement>();