How kinght-code/osintgpt works
OSINT GPT is a Python toolkit for analyzing text with OpenAI models, generating embeddings, finding semantically similar documents, and storing conversation history or vectors. A typical workflow submits text or a query to the package, which can embed it, search a dataframe or Qdrant, and return ranked results; GPT chat interactions can also be persisted in SQLite. The README documents interactive chat and Qdrant collection/vector management. Some source excerpts are partial, so uncertain implementation wiring is omitted.
Main entry points: openai_gpt.py GPT interface · operations.py Semantic operations · openai_embeddings.py Embedding generator · qdrant.py Qdrant adapter
Components 11
Text analysis
- GPT interface
osintgpt/llms/openai_gpt.py - Semantic operations
osintgpt/semantic_operations/operations.py - Embedding generator
osintgpt/embeddings/openai_embeddings.py
Vector search
- Vector engine interface
osintgpt/vector_store/base.py - Qdrant adapter
osintgpt/vector_store/qdrant.py - Document dataframe
Chat and storage
- SQL manager
osintgpt/databases/sql_manager.py - SQLite database
External services
- OpenAI API
- Qdrant service
Other components
- Package user
Connections 12
- Package user → GPT interface: submits prompts
- Package user → Semantic operations: requests search
- Package user → Embedding generator: submits text
- GPT interface → OpenAI API: sends chat requests
- Embedding generator → OpenAI API: requests embeddings
- GPT interface → Embedding generator: embeds query
- GPT interface → Qdrant adapter: searches vectors
- Qdrant adapter → Qdrant service: manages and searches
- Semantic operations → GPT interface: runs similarity searches
- Semantic operations → Document dataframe: searches documents
- GPT interface → SQL manager: persists conversations
- SQL manager → SQLite database: reads and writes