Complete documentation for the SentienceAPI endpoints, request parameters, and response formats
Simple, unified API for all web extraction needs
Reader, Map, and Visual modes for different use cases
Standard HTTP methods, JSON payloads, and clear responses
/v1/observeUniversal endpoint for web content extraction and element coordinate mapping. Supports three modes: Reader (Markdown), Map (Coordinates), and Visual (Screenshot + Coordinates).
Authorization: Bearer YOUR_API_KEY1curl -X POST https://api.sentienceapi.com/v1/observe \
2 -H "Authorization: Bearer YOUR_API_KEY" \
3 -H "Content-Type: application/json" \
4 -d '{
5 "url": "https://news.ycombinator.com",
6 "mode": "read"
7 }'mode: "read"
Extract clean, formatted Markdown content from any webpage. Perfect for LLM context, content analysis, and data extraction.
mode: "map"
Get precise coordinates and metadata for all interactive elements. Ideal for AI agents, automation, and testing.
mode: "visual"
Capture screenshots with element coordinates overlay. Perfect for visual AI models and debugging.
Reduce token costs by up to 95% with intelligent element filtering. Use limit, allowed_roles, allowed_tags, and min_area to focus on what matters.
Choose between Performance (fast, 2 credits) or Precision (pixel-perfect, 10 credits) rendering engines based on your accuracy needs.
Compare engines →All responses include detailed element metadata
1{
2 "engine": "performance",
3 "status": "success",
4 "interactable_elements": [
5 {
6 "id": 1,
7 "uid": "username",
8 "tag": "input",
9 "role": "textbox",
10 "text": "",
11 "selector": "input#username",
12 "bbox": { "x": 100, "y": 150, "w": 300, "h": 40 },
13 "is_visible": true,
14 "z_index": 0,
15 "attributes": {
16 "type": "text",
17 "placeholder": "Enter username",
18 "aria_label": "Username"
19 }
20 }
21 ]
22}400Bad Request
Invalid JSON or missing fields
401Unauthorized
Invalid or missing API key
402Payment Required
Credit limit exceeded
403Forbidden
Account inactive