company research agent

Report Abuse

Basic Information

Agentic Company Researcher is an open source multi-agent application that automates deep company research and produces comprehensive, formatted reports. The repo implements a pipeline of specialized agents that gather and synthesize information from multiple sources such as company websites, news articles, financial reports, and industry analyses. It provides a backend API (FastAPI) and a modern React frontend with real-time progress updates over WebSocket, plus optional Docker deployment and environment configuration for API keys and persistence. The project is structured around research nodes and processing nodes so users can run, inspect, and extend the end-to-end research workflow. The focus is on assembling, filtering, summarizing, and formatting research into downloadable reports rather than on providing a general-purpose agent framework.

Links

Categorization

App Details

Features
The project uses a modular agent pipeline with named research nodes (CompanyAnalyzer, IndustryAnalyzer, FinancialAnalyst, NewsScanner) and processing nodes (Collector, Curator, Briefing, Editor) to organize tasks. It performs multi-source research and Tavily-powered relevance scoring for content curation with a configurable threshold. The platform streams real-time status and report chunks over WebSocket so the frontend can show progress. It uses a dual-model approach: Gemini 2.0 Flash for high-context synthesis and GPT-4.1-mini for final formatting, deduplication, and markdown editing. The stack includes a FastAPI backend, React UI, Docker compose configuration, example environment variable files, and setup scripts to simplify installation and local development.
Use Cases
This repository helps teams and individuals automate time-consuming company diligence by combining data collection, relevance-based filtering, and model-driven summarization into a repeatable pipeline. Users can get structured briefings and a polished report without manual aggregation of sources. Real-time streaming and progress updates improve transparency during long research jobs and the dual-model design balances synthesis of large context with precise formatting. The app supports local development, Dockerized deployment, cloud hosting guides, and optional MongoDB persistence so it can be adapted for one-off research or integrated into existing workflows. The codebase and modular nodes make it possible to customize analyzers, add sources, or tweak curation rules for different use cases.

Please fill the required fields*