Skip to content

brewlens

Bring your brewlytics data into focus.At mission scale.

A model run that ends in a table is a blur of rows. brewlens is the lens: it turns your models' outputs into the application your operators work in — a million detections on a map and a spectrum, a million-edge graph — and every view stays sharp, because the data never leaves the server.

Request a demoSee it at scale
A brewlens application in Present mode, the SBOM and HBOM Risk Console: six stat tiles, findings over time by severity with a disclosure spike, a donut of 66,744 findings by status, a program by severity pivot and a criticality chart

The last mile

A model run is not a mission application

Where most analysis stops

  • Outputs land as tables and files
  • Someone exports, reshapes, and re-hosts them
  • Every new question waits on another hand-built view

With brewlens, the model becomes the app

  • Bind any model output to a map, chart, graph, or table
  • Controls on the page drive the next model run
  • Multi-page, themed, and interactive

The mission just opens it

  • Open a link — no builder, no model knowledge
  • The same sign-in as brewlytics
  • Shared with people and groups by role

How it goes

From model to mission app in three steps

  1. Start

    Start from a mission template

    brewlens comes with working multi-page apps to copy: software and hardware supply chain risk, signals detections, security operations triage, threat hunting, a common operational picture, collection management, and readiness and sustainment.

    • Every template runs on demonstration data the moment you open it
    • Or start blank and drag components onto the canvas
    The hardware supply chain page of the SBOM and HBOM Risk Console template: stat tiles, a world map of facilities coloured by kind, and a donut of lots by provenance

  2. Connect

    Point it at your model and make it yours

    Swap the demonstration query for a brewlytics model or a brewDB source. Its outputs feed the components; the selects, date ranges, and drawn areas on the page become its inputs. Then rearrange, add a drill-down, and pick a theme.

    • Runs on load, on a timer, on demand, or whenever an input changes
    • Counts and sums are computed over the whole result, not a preview
    • Click a bar or brush a range and the whole page filters with it
    The brewlens builder editing the SBOM and HBOM Risk Console: the component palette on the left, the posture page on the canvas, the page inspector on the right, and the data drawer below listing a dependency graph query of 21,860 nodes and 93,280 edges

  3. Publish

    Put it in front of the mission

    Preview runs the app as an operator will see it. Present puts it full-screen on the wall. Publish gives it a link of its own, with the builder gone.

    • Viewer, Editor, and Owner roles for people and groups
    • A classification banner follows your site configuration
    The emitter catalog page of the SIGINT Detections Workbench template in Present mode: stat tiles, a frequency by pulse repetition scatter plot, stacked bars by platform and affiliation, and a category donut

Built for scale

Other app builders fetch the dataset. brewlens never sends it.

Most builders pull rows into the browser and draw what fits. brewlens keeps every result in its data core and answers each component with exactly the slice it draws, so the size of the answer follows the screen, not the dataset — and the picture stays in focus however far you zoom out.

600,000 map positions answered as density, with an exact count
0.48 s
sent to the browser for that map — the same data as points is 2.4 MB
7.5 KB
to bin a one-million-row scatter plot on the server
< 140 ms
laid out and clustered by the data core in about five seconds
1M edges

Measured in brewlens engineering test runs on demonstration data, on a single workstation. Your data and hardware will differ.

The map

A million detections on one map

Zoomed out, the data core answers with density cells and an exact count. Zoom in and the same layer resolves into individual points, lines, and polygons you can click. The answer is sized to the screen, so the map stays responsive as the dataset grows.

  • 600,000 positions returned as 1,740 density cells in under half a second
  • Density, points, and error ellipses layered on one map
  • Draw an area on the map to filter the page or feed a model input
A brewlens map with one million detections in view drawn as a density layer over the western United States, with emitter and sensor points and location ellipses on top, beside stat tiles and a signal-to-noise histogram

The charts

Charts that count every row

Columns, bars, lines, areas, donuts, heat maps, histograms, dot plots, scatter plots, and pivot tables. Each one is an aggregate over the entire result, computed where the data lives — never a sample of whatever reached the browser.

  • A million-point scatter draws as density in well under a second
  • Brush a range or click a mark to filter every component on the page
  • Tables page through the whole result and say how many rows there are
A brewlens page charting one million detections: a time by frequency density plot, bars by band, a donut by emitter category, and an hour by band heat map

The network graph

Graphs past the point where browsers give up

Layout and community detection run in the data core, not in the tab. The browser receives finished positions and draws them, so a graph too large to lay out locally still arrives as a picture you can read.

  • A one-million-edge graph laid out and clustered in about five seconds
  • A 100,000-edge graph ready in about a second and drawn at 60 frames a second
  • Up to 500,000 edges drawn whole; pick a node to pull just its neighbourhood
A brewlens network graph of 21,860 nodes and 93,280 edges grouped into coloured communities, beside a table of packages by reach

Mission templates

Start from an app that already works

Each template is a complete multi-page application on demonstration data. Copy one, point it at your model, and make it yours.

The posture page of the SBOM and HBOM Risk Console template

SBOM & HBOM Risk Console

Findings and triage, the blast radius of one library, a dependency graph, and a hardware supply chain map.

The spectrum page of the SIGINT Detections Workbench template

SIGINT Detections Workbench

A million intercepts on a time × frequency spectrum, geolocated with error ellipses, and an emitter catalog.

The network graph page of the Threat Hunt Workbench template

Threat Hunt Workbench

Indicators by feed and campaign, a per-indicator hunt page, and a network graph of up to a million edges.

Also included: Common Operational Picture · SOC Triage Console · ISR Collection Manager · Readiness & Sustainment · SBOM Intake.

New to the platform? See how a workflow is built on the brewlytics canvas, or where the data lives in brewDB.

See your own data in focus.

Request a demo