VERA documentation¶
VERA packages a source document, parsed structure, embeddings, keyword index,
figures, and citation metadata in one portable .vera file.
Start here¶
- Getting started — install VERA, convert a PDF or Markdown file, and run the first cited search.
- Run the desktop app — install dependencies, start the development app, and create an unpacked build.
- Examples and recipes — copyable workflows for individual documents and document libraries.
- Troubleshooting — installation, conversion, search, indexing, and optional-dependency problems.
User guides¶
- Convert documents
- Search documents
- Search and index document libraries
- Work with figures and highlight regions
- Validate archives and export source documents
- Evaluate retrieval quality
- Use the Python API
- Connect an MCP client
Reference¶
- CLI reference — command overview with links to the exhaustive JSON and exit-code contract.
- API reference — generated Python API docs.
- Current VERA 0.2 format specification
- Legacy VERA 0.1 format specification (deprecated)
- Desktop app product overview
- Collection index design and behavior
- Library index structure diagrams
- Portable Agent Skill
- Agent-skill installation and authoring
Contributor and architecture documentation¶
- Repository architecture
- Desktop app architecture
- Ingest layout lab (vera-lab) — contributor HTML report for pipeline blocks, chunks, and figures
- Additional source formats and visual grounding — Markdown ingest, plugin naming, and remaining format/viewer work
- Changelog
- Roadmap
The README is the product overview, these pages are the human user
documentation, and plugins/vera/skills/vera-search/ is the self-contained package for AI agents.
When behavior changes, update all affected layers in the same change.