Multimodal Retrieval API for AI Agents
Turn scattered visual documents into a queryable knowledge base your agents can reason over —
charts, slides, PDFs, and financial reports extracted, indexed, and cited at scale.
🔍 Extract — VLM-OCR pulls structured facts from charts, tables, slides, and diagrams
🕸️ Index — Ontology graph connects facts across documents by entity, metric, and time
💬 Query — Natural-language questions return cited answers traceable to source and page
| Repo | ||
|---|---|---|
| 🐍 | polyvia-sdk-python | Official Python SDK |
| 📦 | polyvia-sdk-typescript | Official TypeScript / JavaScript SDK |
| 🤖 | skills | Agent skills for Claude Code, Cursor, and other AI clients |
| 📖 | docs | Public documentation source |
| Polyvia API | REST API — ingest documents, manage groups, query with natural language |
| MCP Server | Hosted MCP at app.polyvia.ai/mcp — connect any MCP-compatible AI client in one line |
| Agent Skills | Drop-in skills for agentic coding environments — npx skills add polyvia-ai/skills |
| Polyvia Studio | No-code web app for uploading documents and querying your knowledge base |
polyvia.ai · docs · app · blog · request access
