01 · Dashboards
Dashboards
Composable dashboards built from graph queries. Each widget is a live query plus a visualization spec, so every tile refreshes as the underlying nodes refresh.
Who uses it: Leadership running QBR prep, Ops teams in weekly review, CI analysts shipping executive briefings.
- 20+ widget types · charts, tables, signal feeds, citation streams
- Cross-filter between widgets · click a row, the whole page filters
- Scheduled snapshots, PDF export, shareable links with RLS enforced
- Per-cell provenance · click any value to see source, model, timestamp
02 · Canvases
Canvases
Whiteboard surface for diagramming strategy, mapping the graph, and free-form authoring · nodes from the 88 universal types with typed edges and real-time layout.
Who uses it: Strategy teams running planning workshops, CI analysts mapping landscapes, PMs sketching customer journeys.
- 30 facilitation templates across 10 business domains
- Real-time co-editing with presence cursors and CRDT merging
- 200+ canvas controls · centrality, pathfinding, bundling, layout
- Snapshots, presentation mode, embeds, accessibility audits
03 · Notebooks
Notebooks
JupyterLab on the live graph. Every kernel boots with the `pyramyd` helper preloaded, so any SELECT becomes a DataFrame and cell outputs render PYRAMYD widgets inline.
Who uses it: Data scientists tracing multi-hop questions, ML engineers loading sklearn/xgboost models, PMs prototyping scoring.
- `pyramyd.query()` returns a pandas DataFrame against the live RDS
- Typed handles · `dataset`, `model`, `agent`, `graph`, `dashboard` resolve by name
- Multiplayer via Y.Doc + Hocuspocus · cursor presence, CRDT cell merging
- Auth0 JWT auth · plaintext credentials never live in kernel memory
04 · Sheets
Sheets
Spreadsheet-style live views of graph data. Edit in place; changes are versioned and pushed back to the underlying nodes through the same Smart Action pipeline.
Who uses it: Ops teams and analysts who think in tables · account scoring, ICP builds, feature-market-fit matrices.
- Live joins across node types · no manual VLOOKUP, no stale CSV imports
- Cell-level provenance · every value links back to source and timestamp
- Formulas + APEX cells · a cell can run an arbitrary graph query inline
- Real-time multiplayer editing with presence cursors and CRDT merging
05 · Documents
Documents
Long-form documents · PRDs, RFP responses, battle cards, briefing memos · co-authored with APEX and citation-grounded against the graph at every claim.
Who uses it: Product managers writing PRDs, proposal teams answering RFPs, CI analysts authoring briefings.
- Cite-as-you-type · every claim links to its underlying node
- Smart Action · highlight a sentence, ask APEX to expand, source, or rewrite
- Version control with diff-against-graph · what facts changed since draft?
- Export to Markdown, Word, Google Docs, or PDF with citation footnotes
06 · Slides
Slides
Slide decks built from live graph queries. Charts and tables refresh from the graph at present time, not from a CSV someone pasted in three weeks ago.
Who uses it: Executive briefings, board updates, sales QBRs, CI deep-dives with live refresh on stage.
- Persona templates · CI briefing, RFP debrief, board update
- PowerPoint and Google Slides export with embedded provenance
- Live mode · refresh any chart in-meeting without leaving the deck
- AI-drafted speaker notes grounded in the underlying graph nodes
07 · Schemas
Schemas
Admin surface for the 88 universal nodes, their discriminators, and your tenant's custom fields. Schema changes propagate to every other tab instantly.
Who uses it: Workspace admins, RevOps leadership, and IT teams managing taxonomies and permission models.
- Browse the 88 universal nodes and their principal filter values
- Add custom fields to any node within the type-system constraints
- Permission model · who can read / write / admin each node type
- Audit log of every schema change with one-click rollback
08 · Agents
Agents
Flowise on the live graph. Visual builder where triggers (signal, schedule, manual) flow into actions (query, draft, post, push) · every node a typed graph operation.
Who uses it: Ops teams automating routine work, CI teams running weekly battle-card refresh, RevOps scoring pipeline.
- Full Flowise drag-and-drop builder pre-stocked with graph-aware nodes
- Every run logged with provenance, token cost, and step timing
- Multi-step runs with MCP tools and APEX skills as first-class nodes
- Pause-on-condition guards for human-in-the-loop approval gates
09 · Models
Models
H2O.ai on the live graph. Author, train, evaluate, score, deploy, and monitor ML models · predictive and analytical models surfaced inline next to the entities they describe.
Who uses it: Data scientists, ML engineers, and analytics leaders building forecast, churn-risk, and propensity models.
- H2O AutoML, Driverless AI, Document AI, RAG chat, fine-tuning
- APEX skills for train / predict / explain / deploy in plain English
- Frames imported from files or databases with full transformation lineage
- Model state, deployments, alerts, runs persisted in tenant-isolated storage
