Pecauva: AI financial research platform

Finance teams have to work through dense documents quickly without losing context. Pecauva brings document search, research organization and review into one workspace instead of scattered manual steps.

I built it solo, end to end: the interface, backend workflows, document pipeline, retrieval layer, research matrix and artifact generation. It's currently in testing.

Ask one question and get a document-level answer from every uploaded file, side by side, so you can compare how each report responds.

A dossier for each project, holding research, drafts, playbooks, results and compliance review in one place.

A RAG chatbot that answers only from the project's uploaded documents, cites its sources and stays grounded in the research set.

Ingest: files are organized by source, type and review context.

Retrieve: semantic, document-aware search.

Review: a research matrix for comparing findings and validating answers.

Produce: reusable workflows that generate CSV, PDF and XLSX exports and compliance-oriented drafts.

Next.js, ElysiaJS, PostgreSQL, Qdrant, Redis, object storage, sandboxed analysis infrastructure and export pipelines.

Full case study: https://azizsaadana.dev/work/pecauva