Synthesis Workbench Overview

Synthesis Workbench is a deep literature analysis platform provided by Zotero Agents. It transforms your library into a structured knowledge network, supporting topic synthesis, citation analysis, concept management, and controlled vocabulary management.

How to Open

1. Open Dashboard / Synthesis Workspace via the toolbar button or menu
2. Switch to the Synthesis view in the Workspace Tab

All Surfaces (Pages)

Synthesis Workbench consists of 8 surfaces, each providing a different functional view:

Surface	Function	Docs
Home	Library overview dashboard: library insights (registered papers / topic count / graph nodes), Git sync status panel, trending topic card list	Details
Topics	Topic list and management: 3 view modes (graph / grid / list), create and update topics, topic search and sorting	Details
Index	Canonical Reference index: paper registry view (paper list + citation rows + binding status), canonical reference view (search / merge / redirect / deduplicate)	Details
Review	Review hub: 3 sub-tabs — citation match review (accept/reject binding proposals), concept review, topic graph relationship review	Details
Graph	Citation graph visualization (force-directed / radial / component — 3 layouts), with topic filtering and node/edge interaction	Details
Tags	Controlled tag vocabulary management + automatic tagging suggestion approval	Details
Concepts	Concept knowledge base management: four-layer structure of concepts / senses / aliases / relations, overlayable onto the topic graph and reader	Details
Reader	Topic reader: full Topic Detail page with 8 sub-pages (Overview, Taxonomy, Claims, Compare, Future Directions, Coverage, References, Report)	Details

Core Concepts

Canonical Store

The Canonical Store is the underlying knowledge graph storage for the Synthesis system. It stores content-addressable JSON files in the Zotero data directory.

Storage location: <Zotero data directory>/zotero-agents/data/synthesis/

Directory structure:

synthesis/
├── topics/             # Structured artifacts for topic synthesis
├── concepts/           # Concept knowledge base
├── topic-graph/        # Topic graph nodes and edges
├── citation-graph/     # Citation graph snapshots
├── tags/               # Controlled tag vocabulary
├── sync/               # Git sync working tree
└── state/              # Runtime state (transactions, receipts, caches, etc.)

Each file uses a JSON envelope format (CanonicalEnvelope) that includes a schema ID, version number, timestamp, and schema-validated data body. Write operations use transactional semantics: data is first staged in the transaction directory, promoted to the canonical location upon successful validation, and automatically rolled back on failure.

Reference Sidecar

A Reference Sidecar is an index of the attached artifacts for each paper. When a workflow processes a literature item and generates a digest, reference list, and citation analysis, these artifacts are attached to the item as structured notes (Zotero Notes). The Sidecar system scans these notes and records artifact status (complete / partial / missing) into the index.

Sidecar scan cycle: The sidecar is triggered to scan at the following times:

• After a workflow execution completes and writes artifacts
• When an explicit sidecar refresh operation is triggered
• When the system detects stale sidecar data at startup

Artifact types:

Artifact	Description
digest	Paper digest (Markdown)
references	Reference list (JSON)
citation_analysis	Citation analysis report (JSON)

Sidecar data serves as the primary input to the Canonical Reference Index — the system extracts citation records from the references artifact, establishes canonical references, and then attempts to match and bind them to library items.

Data Flow

Zotero Library
    │
    ├──→ Workflow Execution (Literature Analysis / Deep Reading)
    │         │
    │         ↓
    │   Artifact Notes (Digest / References / Citation Analysis)
    │         │
    │         ↓
    │   Reference Sidecar ← Scan artifact status
    │         │
    │         ├──→ Canonical Reference Index
    │         │         │
    │         │         ├──→ Citation Binding (Bind to Zotero Items)
    │         │         └──→ Citation Graph
    │         │
    │         └──→ Topic Synthesis
    │                   │
    │                   ├──→ Topic Graph (Topic Relationships)
    │                   └──→ Concept Associations (Concept KB)
    │
    └──→ Git Sync ←→ Remote Repository (Version Control and Backup)

Prerequisites

Using Synthesis Workbench requires:

• A configured Skill-Runner backend (for running synthesis workflows)
• Paper items already present in the library

Next Steps

• Home Dashboard — View library overview and sync status
• Tags Management — Manage the controlled tag vocabulary
• Index & Citation Graph — Learn about reference indexing and citation networks
• Create Topic Synthesis — Create topic analyses
• Review Hub — Review citation matches, concepts, and topic graph proposals
• Concept Knowledge Base — Manage core concepts
• Git Sync — Configure data sync and backup