How sourabh160705/satquery-ai works
SatQuery AI is a satellite-imagery analysis platform: users select or acquire imagery, submit natural-language questions, and receive analysis with visual evidence and an auditable trace. The README documents single-image VQA, captioning, grounding, bi-temporal change detection, optical–SAR fusion, spectral analysis, and GeoTIFF ingestion. The Python API and agent workflow are supported by sampled source; some location and integration behavior is documented or visible only in the Node server excerpt, so uncertain connections are omitted.
Main entry points: orchestrator.py Query Orchestrator · main.py Python API · Studio.jsx Studio Workbench · app.js Node API
Components 18
Analysis Workbench
- Studio Workbench
satquery-ai/client/src/components/Studio.jsx - Client Shell
satquery-ai/client/src/main.jsx
API and Imagery
- Node API
satquery-ai/server/app.js - Python API
backend/main.py - Image Parser
backend/engine/geotiff_parser.py - Location Imagery
backend/engine/location_satellite.py
Agent Workflow
- Query Orchestrator
backend/agent/orchestrator.py - Intent Classifier
backend/agent/intent_classifier.py - Execution Trace
backend/agent/tracer.py
Analysis Tools
- Image VQA
backend/specialists/vqa_specialist.py - Scene Captioning
backend/specialists/captioning_specialist.py - Region Grounding
backend/specialists/grounding_specialist.py - Change Analysis
backend/specialists/change_specialist.py - Optical SAR Fusion
backend/engine/optical_sar_fusion.py - Spectral Indices
backend/engine/spectral_indices.py
Other components
- Analyst
- Geocoding Service
- Satellite Imagery Source
Connections 20
- Analyst → Studio Workbench: submits query
- Studio Workbench → Client Shell: renders in
- Studio Workbench → Python API: requests analysis
- Studio Workbench → Node API: requests location data
- Node API → Geocoding Service: geocodes places
- Node API → Satellite Imagery Source: acquires imagery
- Python API → Image Parser: parses uploads
- Python API → Location Imagery: fetches scenes
- Location Imagery → Satellite Imagery Source: retrieves scenes
- Python API → Query Orchestrator: processes queries
- Python API → Spectral Indices: computes indices
- Query Orchestrator → Intent Classifier: classifies intent
- Query Orchestrator → Image VQA: answers questions
- Query Orchestrator → Scene Captioning: describes scenes
- Query Orchestrator → Region Grounding: localizes regions
- Query Orchestrator → Change Analysis: analyzes changes
- Query Orchestrator → Optical SAR Fusion: fuses modalities
- Query Orchestrator → Spectral Indices: computes indices
- Query Orchestrator → Execution Trace: records execution
- Scene Captioning → Spectral Indices: computes NDVI