OSINT GPT is an open-source intelligence analysis tool hosted on GitHub, with about 530 stars under the Apache-2.0 license and updates in 2026. Instead of searching the open web, it indexes documents an analyst already holds and answers questions from that material with citations, so every answer keeps a path back to its source.
Analysts organize material into projects, add documents, build an index, and then ask questions from the command line or through the Python library. Work stays separated per project, which suits investigations where collections must not be mixed. The tool is provider-flexible: embeddings can come from OpenAI, Gemini, Voyage, Ollama, or local sentence-transformers models, while answer generation supports OpenAI, Gemini, Anthropic, and Ollama. Storage backends include SQLite, Qdrant, and Postgres, and a fully local configuration keeps sensitive material on your own machine.
The day-to-day workflow is deliberately simple: create a project, point the tool at a folder of material, build the index, and start asking questions. Because embedding and generation providers are interchangeable, an analyst can begin with a hosted model for convenience and later move the same collection to fully local models when a case turns sensitive, without reorganizing the underlying material or giving up cited answers.
OSINT GPT is free and open source under the Apache-2.0 license. You only pay for any hosted model APIs you choose to plug in; a fully local setup with Ollama and sentence-transformers runs without provider fees.
It suits OSINT analysts, investigators, journalists, and researchers who need to interrogate large document collections they already possess, and who care about cited, source-traceable answers rather than free-form chat.