A provider-agnostic, multi-agent financial research crew built on CrewAI.
Point it at a ticker and a research question; a team of specialised agents researches the company, pulls market data, reads filings, assesses risk, and writes a cited investment brief.
Serious financial research is a pipeline of distinct skills: gathering sources, crunching numbers, reading disclosures, weighing risk, and communicating clearly. finsight-crew models each as a focused agent and lets CrewAI orchestrate the hand-offs.
pip install -e .
cp .env.example .env # add your API key
finsight research AAPL --question "Is the balance sheet healthy?"Early development. See CHANGELOG.md for progress.
See docs/usage.md and docs/agents.md.
- Provider-agnostic LLM (Claude default, OpenAI/Ollama swappable)
- Tools: web search, market data, SEC filings, RAG, calculator, sentiment
- Markdown / JSON / HTML briefs, multi-ticker comparison
- Report evaluation and run observability