Target
Start with a lawful scope, research question, and limits before any automation runs.
OSINT Agents
This section collects open-source projects around OSINT agents, MCP servers, LLM-based investigation workflows, and research automation. The goal is not to list every tool, but to keep a focused set of repositories that are useful for studying how agentic OSINT systems can be designed.
Reference architecture
A practical copilot should keep the reasoning loop separate from deterministic tools, store evidence as case state, and route sensitive conclusions through human review before reporting.
Investigation workflow
Start with a lawful scope, research question, and limits before any automation runs.
Use MCP servers, search tools, CTI feeds, archives, and browser agents for repeatable collection.
Normalize names, accounts, domains, observables, aliases, and documents before linking them.
Move findings into graph or case-memory systems so links can be reviewed instead of guessed.
Separate facts, signals, inferences, and uncertainty. Do not let the model decide alone.
Use agents to propose lines of inquiry, then require sources and contradiction checks.
Turn the trace into a clear report with citations, confidence, limits, and next steps.
28 repositories shown
Metadata is a manual snapshot from GitHub on 2026-06-07. Re-check stars, license, and activity before procurement or production use.
OpenOSINT/OpenOSINT
AI-powered OSINT agent with CLI, REPL, MCP server, web UI, and modular OSINT tools.
Useful for
Useful for studying an end-to-end OSINT agent that lets an LLM request real tool output instead of inventing results.
OSINT relevance
Maps directly to username, email, breach, WHOIS, DNS, Shodan, VirusTotal, GitHub, and reporting workflows.
soxoj/awesome-osint-mcp-servers
Curated directory of MCP servers that expose OSINT tools and services to LLM clients.
Useful for
Useful as a discovery map before choosing which MCP servers to connect to an investigation copilot.
OSINT relevance
Helps compare username, search, social, domain, and threat-intelligence MCP options in one place.
w0h1v/mcp-maigret
MCP wrapper for Maigret username checks across public profile sources.
Useful for
Useful for controlled username research where public-source matches need to be reviewed by a human.
OSINT relevance
Fits the entity discovery phase, but results require careful false-positive handling.
frishtik/osint-tools-mcp-server
Python MCP server exposing multiple OSINT tools to assistants such as Claude.
Useful for
Useful as an example of packaging several existing OSINT utilities behind one MCP interface.
OSINT relevance
Shows the design tradeoff between convenience and the need for strict scope, rate, and safety controls.
w0h1v/mcp-shodan
MCP server for Shodan lookups, DNS operations, CVE context, and internet-exposure review.
Useful for
Useful for passive infrastructure visibility checks inside an agent workflow.
OSINT relevance
Supports domain, IP, exposed-service, and vulnerability context without adding active scanning.
damionrashford/RivalSearchMCP
FastMCP research server for web search, social search, academic databases, news, entities, and documents.
Useful for
Useful for studying deterministic tool output for agent chaining without an LLM embedded inside the server.
OSINT relevance
Maps to collection, entity profiling, conflict detection, and document review stages.
assafelovic/gpt-researcher
Autonomous deep-research agent for web and local research across multiple LLM providers.
Useful for
Useful for studying source gathering, synthesis, and report generation loops.
OSINT relevance
Maps to collection, source comparison, evidence summarization, and report draft generation.
dzhng/deep-research
Small TypeScript deep-research agent that iterates search, scraping, reasoning, and markdown reporting.
Useful for
Useful as a simple reference implementation before building a heavier investigation copilot.
OSINT relevance
Shows breadth/depth controls, query generation, source processing, and final report generation.
langchain-ai/open_deep_research
Configurable open-source deep research agent across model providers, search tools, and MCP servers.
Useful for
Useful for studying a modern research agent that can swap search and model providers.
OSINT relevance
Good architectural reference for iterative collection, source selection, and report assembly.
SkyworkAI/DeepResearchAgent
Hierarchical multi-agent deep-research system with planning and specialized lower-level agents.
Useful for
Useful for studying multi-agent decomposition and resource lifecycle management.
OSINT relevance
Maps to investigation planning, tool selection, evidence gathering, and iterative refinement.
Yaogui415/newsflow-oss
AI-assisted newsroom workflow for event tracking, evidence organization, approval chains, and corrections.
Useful for
Useful for human-in-the-loop investigation workflow design.
OSINT relevance
Focuses on evidence structuring, review gates, approval traceability, and correction handling.
BharathxD/ClaimeAI
LangGraph fact-checking system that extracts claims, searches evidence, and generates verification reports.
Useful for
Useful for claim-level verification rather than asking an LLM for a single verdict.
OSINT relevance
Maps well to evidence verification, misinformation review, and report-ready claim breakdowns.
MISP/MISP
Open-source threat-intelligence and sharing platform for events, attributes, and communities.
Useful for
Useful as a structured evidence and indicator-sharing backend for defensive investigations.
OSINT relevance
Gives an OSINT agent a mature event/IOC data model instead of a flat note store.
IntelOwlProject/IntelOwl
Open-source platform for threat-intelligence automation, analyzers, connectors, and enrichment jobs.
Useful for
Useful as a backend model for analyzer orchestration and repeatable enrichment pipelines.
OSINT relevance
Connects services such as VirusTotal, Abuse.ch, GreyNoise, URLScan, Shodan, OTX, MISP, and OpenCTI.
4R9UN/fastmcp-threatintel
FastMCP threat-intelligence server for IP, domain, URL, and hash analysis.
Useful for
Useful for adding defensive CTI enrichment to an investigation copilot.
OSINT relevance
Supports IOC context, indicator review, and report enrichment for defensive investigations.
CooperCyberCoffee/opencti_mcp_server
MCP interface for querying OpenCTI threat intelligence through Claude Desktop.
Useful for
Useful for connecting a local or private OpenCTI knowledge base to an assistant.
OSINT relevance
Turns stored observables, reports, and relationships into queryable investigation context.
bornpresident/MISP-MCP-SERVER
MCP server that integrates MISP threat-intelligence data with LLM clients.
Useful for
Useful for teams that already store events, attributes, and IOCs in MISP.
OSINT relevance
Can expose structured indicator and event context to an analyst copilot.
aplaceforallmystuff/mcp-threatintel
MCP server that unifies AlienVault OTX, AbuseIPDB, GreyNoise, and abuse.ch sources.
Useful for
Useful when an agent needs consistent CTI lookups without opening several browser tabs.
OSINT relevance
Helps with source collection, IOC triage, and confidence-building across multiple feeds.
microsoft/graphrag
Graph-based retrieval-augmented generation system for reasoning over narrative private data.
Useful for
Useful for turning investigation corpora into graph memory for question answering and synthesis.
OSINT relevance
Provides architectural ideas for entity extraction, relationship storage, and graph-guided summaries.
OpenCTI-Platform/opencti
Open cyber threat-intelligence platform for structuring, storing, organizing, and visualizing CTI.
Useful for
Useful as the graph-backed case memory behind an investigation copilot.
OSINT relevance
Uses STIX-style relationships and observables that map well to entity and relationship analysis.
neo4j-labs/llm-graph-builder
Builds Neo4j knowledge graphs from PDFs, documents, YouTube videos, web pages, and other unstructured data.
Useful for
Useful for converting collected OSINT material into entities and relationships.
OSINT relevance
Fits document intake, entity extraction, relationship analysis, and graph exploration.
rolandpg/zettelforge
Agentic CTI memory with STIX knowledge graphs, threat-actor alias resolution, offline RAG, and MCP support.
Useful for
Useful for studying memory design around CTI entities, aliases, and offline investigation context.
OSINT relevance
Targets threat-actor alias resolution and graph-backed memory, both useful in entity-heavy investigations.
firecrawl/firecrawl-mcp-server
Official Firecrawl MCP server for web search, scraping, and page extraction.
Useful for
Useful for turning public pages into structured material that an agent can cite or summarize.
OSINT relevance
Supports evidence collection and source text extraction when terms and robots rules allow it.
exa-labs/exa-mcp-server
MCP server for Exa web search and crawling.
Useful for
Useful for semantic web discovery and source collection inside an agent.
OSINT relevance
Can help find sources, similar pages, and context for public web research.
browserbase/mcp-server-browserbase
MCP server that lets LLMs control a browser through Browserbase and Stagehand.
Useful for
Useful when a research agent needs browser automation for public pages that are hard to fetch directly.
OSINT relevance
Fits controlled web collection and screenshot workflows, with explicit human oversight.
tavily-ai/tavily-mcp
MCP server for real-time search, extraction, mapping, and crawling.
Useful for
Useful as a web-search and extraction layer for research agents.
OSINT relevance
Supports collection and source discovery, but conclusions still need human verification.
apify/apify-mcp-server
MCP server that connects agents to Apify actors for search, maps, social, and website extraction tasks.
Useful for
Useful for studying actor-based collection pipelines and reusable extraction jobs.
OSINT relevance
Can support public-data collection, but every actor needs legal, terms-of-service, and scope review.
elb-pr/sleuth
Six-phase OSINT workflow with templates for direction, collection, entity resolution, reasoning, and reports.
Useful for
Useful for studying investigation gates and task files rather than only autonomous tool calls.
OSINT relevance
Closely mirrors target, collection, entity resolution, relationship analysis, hypothesis, and reporting stages.