Skip to content

Roadmap

Three sections, one question each:

  • Shipped

    Released and documented. Use it today.

  • In progress

    Open pull requests, grouped by theme.

  • Ideas

    Wanted, not scheduled. No commitment.

How to read this page

Every dated claim below is traceable to the changelog, which is also where you will find the current release. Dates are deliberately absent from In progress and Ideas: the pull request is the honest status, an estimate would not be.

Big Picture

Foundations

→ v1.0.0 ✅

Dashboards, YAML-defined projects, Delta Lake ingestion, MultiQC and specialized components

React viewer

v1.0.0 ✅

Vite + Mantine SPA becomes the sole frontend on canonical URLs; Dash removed

Live data & access

v1.1.x ✅

Real-time dashboard events over WebSocket, magic-link login, pipeline provisioning

Operability

v1.2.x ✅

Admin Log & Task monitoring, per-phase ingestion detail, pinned timeline footer

Performance

v1.3.0 ✅

Lazy panel loading, bounded payloads, query pushdown, benchmark harness

Dashboard structure

v1.4.0 ✅

Dashboard sections, a dashboard-wide map panel, ten new card layouts

Cross-tab structure & selections

v1.6.0 – v1.7.0 ✅

Sections and filters that live on every tab, named selection groups, the Analysis panel, funnel filtering

Identity & exploration

v1.8.x ✅

Instance and dashboard brand themes, a navigable phylogeny, run provenance from the pipeline

Authoring & the catalog

v1.9.0 ✅

In-app catalog picker, Tool Studio for contributing a tool, use: provenance on a component

Backup & restore

v1.9.0 ✅

Backups created, validated and restored from the admin panel, scheduled, with a retention policy

Pipeline-triggered ingestion

v1.10.0 ✅

A Nextflow pipeline ingests its own results when it finishes

Versioning & the ingestion watcher

v1.15.0 ✅

Dataset and dashboard versioning, time travel, a watcher that re-ingests on its own

Data where it lives, dashboards from a run

v1.16.0 ✅

Run folders read on S3, projects created from a folder in the web UI, and a dashboard composed from whatever the catalog recognises

Project authoring & embedding

In progress 🚧

Project builder, a component embedded in an external site

Ideas

Unscheduled 💡

Serverless Depictio, journeys through tabs, citable DOI snapshots, AI-assisted analysis

Shipped

One row per capability, newest first. The feature docs are the reference; this table only says when something arrived.

Capability Since Docs
A dashboard composed from what the tools catalog recognises in a results directory, the way MultiQC builds its report v1.16.0 Composed dashboards
Data read where it is: an s3:// run folder, a URL, a bucket prefix or a data manifest; projects created from a run folder in the web UI, refreshed in place, exported as a template v1.16.0 Remote data and manifests · Export a project as a template
Dashboard version history: bookmarks, read-only previews, restoring a dashboard or one component v1.15.0 Dashboard and data versions · Component history
Every write a Delta commit; dashboards, previews and datasets read as a past commit wrote them v1.15.0 Data versions
depictio watch re-ingests on its own, with a Watchers pane and live ingestion steps v1.15.0 Watch Command · Monitoring
Ingestion history on the project page, and an ingestion started from the browser v1.15.0 Ingestion history · Run ingestion
Landing tabs: block markdown, tab links, filter bars, minimal figures, a phylogeny summary v1.14.0 Landing tabs
A built-in Guide on every dashboard v1.14.0 The dashboard Guide
Search every component of every tab with ⌘K v1.14.0 Spotlight search
Comments and annotations on dashboard components, with agent proposals held for review v1.13.0 Comments and Annotations
Thirteen more nf-core templates, ten advanced visualization kinds including a GenomeSpy genome view v1.13.0 Pipeline Templates · Components
Genome regions carried between collections, record cards that follow a selection v1.13.0 Cross-DC Filtering · Interactive Selection Filtering
The full server installed as a Python package, no containers: depictio local up v1.12.0 Python package
One depictio package, depictio ingest <results dir>, refreshes that keep edited dashboards v1.12.0 Ingest Command · Refreshing a project
Recipes declare the schema of their inputs and their output, checked around the transform v1.12.0 Recipes
SeaweedFS as the bundled S3 store, replacing MinIO v1.12.0 Upgrading to v1.12.0
An opt-in feedback link that carries the dashboard and tab v1.11.0 Feedback
The funnel overview follows the values of one column v1.11.0 Funnel filtering
Ingestion triggered by the pipeline itself, when it completes v1.10.0 Nextflow trigger
A run directory identifies its own pipeline, so a template needs no flag v1.10.0 Templates
nf-core/variantbenchmarking template and benchmarking viz kinds v1.10.0 variantbenchmarking
Backups created, scheduled and restored from the admin panel v1.9.0 Backup & Restore
Pick a component from the catalog, with use: provenance on it v1.9.0 Picking from the catalog
Tool Studio: contribute a catalog tool from the browser v1.9.0 Tool Studio
Run provenance, a real tab header, per-section grid packing v1.8.3 Templates · Dashboards
Password sign-in disabled, shared code for public mode v1.8.2 Authentication Modes
Instance and dashboard brand themes, a navigable phylogeny v1.8.0 Branding · Components
Selection groups, the Analysis panel, funnel filtering v1.7.0 Interactive Selection Filtering · Dashboards
Sections and filters that span every tab of a dashboard v1.6.0 Dashboards · YAML Sync
Filters carried into the component builder, bucket-scoped S3 startup v1.5.2 Dashboard creation
Cross-version backup compatibility guard v1.5.2 Backup & Restore
Anonymous installation telemetry v1.5.0 Telemetry
Dashboard sections, dashboard-wide map panel, ten new card layouts v1.4.0 Dashboards · Components
Lazy panel loading, bounded payloads, Load all, benchmark harness v1.3.0 Performance & Scaling
Pinned timeline footer, compact tables v1.2.2 YAML Sync
Admin Log & Task monitoring, per-phase ingestion detail, run provenance v1.2.0 – v1.2.1 Monitoring
Real-time dashboard events over WebSocket v1.1.4 Real-time Events
Passwordless magic-link login, pipeline provisioning v1.1.3 Authentication Modes
Cross-DC linking UI, ingestion reports, self-adapting nf-core templates v1.1.0 Cross-DC Filtering
React viewer as sole frontend on canonical URLs v1.0.0 Dashboards
Tools catalog — tool outputs become dashboard components v1.0.0 Catalog
Advanced biology visualizations (volcano, Manhattan, sunburst…) v0.13.0 Components
MultiQC data lifecycle, DC-level type config v0.12.0 Projects Guide
Templates & recipes — one-command project setup v0.10.0 Templates · Recipes
Geospatial maps, choropleth & GeoJSON collections v0.8.0 Components
MultiQC report integration v0.5.0 Components
YAML-defined projects & dashboards, Delta Lake ingestion via CLI v0.0.5 CLI Usage · YAML Sync
Docker Compose & Kubernetes/Helm deployment v0.0.5 Installation

In progress

Grouped by theme. Each item links the pull request that implements it, so its real state is always one click away.

Authoring: templates and project setup

Lowering the cost of contributing a tool or standing up a project. The catalog picker and Tool Studio both landed in v1.9.0, and a dashboard composed from a results folder in v1.16.0; what is left is below.

  • nf-core template harmonization: the bundled templates reference the same catalog entries a dashboard does, rather than carrying bespoke inline tiles (#873)
  • Project builder: depictio project-builder <folder> turns a folder into a project.yaml, with live glob/regex matching and schema-consistency checks (#901)

Component export & embedding

  • Embed a component in an external site: serve one dashboard component either as a Plotly spec for your own plotly.js, or as a single self-contained offline page. Off by default (#917)

Ideas

Wanted, discussed, not scheduled. Nothing here has a date, and some of it will never be built. Open or upvote an issue to push something up.

Idea What it would give you Link
Serverless Depictio Explore a dashboard with no server to deploy or maintain: S3 files and prerenders read client-side, with WASM and DuckDB #934
Journeys Named paths through tabs. The two halves of this idea already shipped: filters promoted to dashboard scope in v1.6.0, and funnel filtering in v1.7.0 #756 · #690 · #691
Citable science DOI-backed snapshots per dashboard version, sample-to-viz provenance via LabID, static export to Quarto. Nothing is open for it yet; #931 is the nearest neighbour, surfacing pipeline and filtering provenance in a dashboard #931
AI-assisted analysis Describe the analysis and get a proposed layout; anomaly detection, narration, and an MCP server for AI agents #844 · #79
Grain-aware joins Detect each table's grain so a cross-DC join cannot silently explode rows #877 · #876
Snakemake report plugin Depictio as a drop-in replacement for Snakemake's static HTML report none yet

Reproducibility & FAIR

Depictio is built with the FAIR principles in mind. Most of the reproducibility story is shipped; citability is not.

Challenge Status
Reproducing a setup takes manual work nf-core templates, one-command project setup (docs)
A visualization can't be reproduced YAML-defined dashboards over traceable data (docs)
Experiment data sits in silos Cross-DC linking (docs)
Dashboards disappear Hosted on SciLifeLab Serve
No data lineage Run provenance records the parameters a run used; each table write is a Delta commit that a dashboard can be read at (docs)
Not citable, no sample-to-viz traceability DOI snapshots + LabID provenance — an idea

Get involved