Platform · Overview · v0.1.0-beta.1

One data layer. Several ways to model data.

Structured rows, documents, time-ordered events and embedding vectors in one system — one planning and execution path across all of them. Graph and blobs extend that same layer, and are on the roadmap.

plomid · five layers, one system explore

Turns a request into a plan over one data layer.

  • One plan across data models
  • Predicate and layout awareness
  • Consistent semantics
Requests enter at the top, land on storage The full stack ↓
The premise

Most teams end up with four databases and a pipeline.

Each system is good at its own shape of data. The cost shows up in the seams: the code that keeps them consistent, and the questions neither one answers alone.

The journey

Follow one question.

Pick one of the questions platform teams carry, then read the same question twice: first the estate it has to travel today — stores, pipelines and the seams between them — and then as one request, in one system.

What is happening now, to which record, and what led here.

Where it goes today

  1. 01
    Event broker

    Events land in a queue first — partitions, replay and retention, a system with its own on-call.

  2. 02
    Time-series store

    The signals themselves: rollups, downsampling and a retention policy of their own.

  3. 03
    Application database

    The record each event is about — the join key lives here, not with the events.

  4. 04
    Alerting pipeline

    A bespoke consumer correlates the window, the record and the threshold — written and owned by your team.

Window from one store, identity from another, correlation from a service you own — late events and retention gaps surface as wrong alerts.

Three systems · one pipeline

Where it goes in PLOMID

One request

Telemetry and events are stored with the records and documents they describe, so a window, the entity behind it and its context are read from one layer.

  • Events and measurements
  • Windowed reads with the record attached
  • No export between systems
Reads, in one request
  • Time series
  • SQL
  • JSON

One plan · one storage contract · no exports between stores

One record, read whole: its rows, its documents, and the history behind them.

Where it goes today

  1. 01
    Application database

    Rows, keys and joins — the record that fits columns.

  2. 02
    Document store

    The fields that keep changing shape, in a second deployment with its own backup.

  3. 03
    Audit / history store

    Every change written by application code or a change stream — the copy nobody wants to query.

  4. 04
    Profile service

    The “whole record” view fans out to all three on every read, and owns every failure mode.

Three stores kept in step by application code: the whole record is stitched together per request, and every new field means touching all of them.

Three systems · one seam

Where it goes in PLOMID

One request

One request reads rows, keys and joins alongside nested documents and the change history behind them, instead of loading one side into the other.

  • Rows for core records
  • Documents for changing shape
  • History read beside both
Reads, in one request
  • SQL
  • JSON
  • Time series

One plan · one storage contract · no exports between stores

Today's number, asked of live operations — filtered by the fields that don't fit columns.

Where it goes today

  1. 01
    Operational database

    The live numbers, serving the application.

  2. 02
    Change data capture

    Changes streamed out of the database — a pipeline with its own schema-evolution problems.

  3. 03
    Warehouse load

    Staging, deduplication and backfills — jobs that run on a schedule and run late after every failure.

  4. 04
    Reporting warehouse

    The copy the dashboards read — as of its last successful load.

  5. 05
    Metric layer

    The number defined a third time, in model files the application never sees.

The report describes the warehouse as of its last load, the metric layer defines the number again — the argument between systems is standing.

Two systems · two pipelines

Where it goes in PLOMID

One request

Aggregates and windows run over the live tables with document fields filterable beside them, so the number you report and the number you serve come from one place.

  • Aggregates in place
  • Document fields filterable
  • One copy of the truth
Reads, in one request
  • SQL
  • JSON
  • Time series

One plan · one storage contract · no exports between stores

Find what is similar — ranked, filtered by structure, and current as of the last write.

Where it goes today

  1. 01
    Application database

    The records themselves — the source of truth.

  2. 02
    Embedding pipeline

    Vectors computed as rows change — a model, a service and a queue between source and search.

  3. 03
    Vector store

    A separate index: the vectors and their metadata are both copies of the records.

  4. 04
    Re-embedding jobs

    A model upgrade means backfilling the whole corpus — a migration with no downtime window.

Every change has to be pushed through the pipeline or the index silently lags — deletes leave orphans, results describe records that moved on, and a model upgrade means re-embedding everything.

Two systems · two pipelines

Where it goes in PLOMID

One request

Embeddings are stored beside the records and documents they describe and ranked with distance ordering, filtered by structure, so similarity is an access path over the data that already serves the application.

  • Dense vector columns
  • Exact top-k, HNSW at scale
  • Metadata filtering beside ranking
Reads, in one request
  • Vector
  • SQL
  • JSON

One plan · one storage contract · no exports between stores

Note The “today” picture is the multi-store pattern the premise above describes — four databases, joined by a pipeline. The PLOMID picture states only what the product does today; the model explorer below opens up what each shape is made of. Open the model explorer

Model explorer

Explore the models.
Four you can use today.

Pick a model to see the shape of the data it holds. Each one is a way of shaping the same underlying data, the dashed ones extend the layer rather than replacing it.

Solid, in the product Dashed, still being built

SQL Transactional application data join · plan

Structured queries, down to the page.

PLOMID · SQL join · plan hover or tap the diagram
one access path: keys narrow before rows are read

Hover the diagram or the text, each answers the other. Click to pin.

Relational workloads, tables, joins, transactions.

Structured data with a familiar query surface. Tables, indexes, relationships and transactions on one engine.

Workloads it carries
JSON Application documents resolve path

Documents without leaving the data layer.

PLOMID · JSON resolve path hover or tap the diagram
one document, addressed branch by branch

Hover the diagram or the text, each answers the other. Click to pin.

Documents and nested objects, queried alongside rows.

Flexible documents live next to structured data, so one query can mix fields from both without a second system.

Workloads it carries
Key-Value Direct access get · set

Direct access by key.

PLOMID · Key-Value get · set hover or tap the diagram
one hop: key → rail → value

Direction, not a capability, hover to explore the shape it will take.

Short-lived state, addressed by key.

One key resolves to one value without a scan, sessions, flags and cached state read as directly as they are written.

Direction
Time series Telemetry and events window scan

Data where time is part of the query.

PLOMID · Time series window scan scrub the plot · switch range

Hover the diagram or the text, each answers the other. Click to pin.

Events, measurements and history over time.

Time-ordered data is stored with the rest of the application data, so history and current state agree.

Workloads it carries
Geo Spatial data radius probe

Location as a queryable dimension.

PLOMID · Geo radius probe move the point · layers, radius, zoom

Direction, not a capability, hover to explore the shape it will take.

Where it is, asked like any other predicate.

Position, distance and range become part of the same query surface, nearby objects answered without exporting coordinates to a second system.

Direction
Vector Search and retrieval similarity probe

Similarity as a native data operation.

PLOMID · Vector similarity probe move the query · k, filters

Hover the diagram or the text, each answers the other. Click to pin.

Embeddings and nearest-neighbour retrieval over stored data.

Dense vectors live in typed columns beside the records they describe, ranked with distance operators and top-k ordering, exact by default with HNSW acceleration at scale.

Workloads it carries
Graph Connected data traverse

Relationships as first-class data.

PLOMID · Graph traverse hover or tap the diagram
one traversal: a node lights its edges

Direction, not a capability, hover to explore the shape it will take.

Nodes, edges and traversal.

Relationships expressed as first-class structures instead of reconstructed at query time.

Direction
Objects Large objects and assets put · get

Binary data with its metadata.

PLOMID · Objects put · get hover or tap the diagram
one write: object → metadata → storage

Direction, not a capability, hover to explore the shape it will take.

Large objects and file-like assets.

Big immutable objects addressed like data, with metadata that stays queryable.

Direction
Hybrid Mixed workloads one request

One request. Several shapes of data.

PLOMID · Hybrid one request toggle the shapes · one request

Direction, not a capability, hover to explore the shape it will take.

Mix models in a single statement.

Rows, distance, a time window, a traversal and an object can be named by one request, with the plan working across the layer that holds them. Mixing models in one statement is on the way, the pieces it needs are already in the layer.

Direction
How the pieces fit

One database. Many ways in. One shared data foundation.

PLOMID brings multiple data models and workloads into a unified database, while applications and agents interact through the interfaces that fit their needs.

ERP ANALYTICS CRM IOT AI APP AGENT SQL DOCUMENT TIME SERIES SDK API · TOOLS PLOMID UNIFIED DATABASE SQL JSON TIME SERIES KEY-VALUE VECTOR GRAPH OBJECTS GEO ROADMAP MONGO WIRE GRAPHQL KV · GEO ACCESS SERVES ANYTHING ANALYTICS BI AI CONTEXT APPS AGENTS DEPLOY ANYWHERE DOCKER SELF-HOSTED CLOUD EDGE

One database. Many ways in. One shared data foundation.

  1. Applications

    AI · CRM · ERP · Agents · IoT · Analytics

  2. Protocols

    SQL · Document · Time series · SDK — Mongo wire and GraphQL on the roadmap

  3. PLOMID — Unified database

    The data models live inside it, not beside it.

  4. Data models

    SQL · JSON · Time series · Vector · Graph · Objects — key-value and geo as roadmap

  5. Serves anything

    Analytics · BI · AI context · apps · agents — all reading the same state.

  6. Deploy anywhere

    Docker and self-hosted today — cloud and edge on the roadmap.

Workloads

Workloads, not industries.

Described by what you are doing with the data, not which sector you are in. Select one — the panel answers in place.

SQL + JSON + Time series Application data

Records, keys and joins with nested documents beside them, one request that reads both.

  • Rows for core records
  • Documents for changing shape
  • One statement across both
Lifecycle

What happens to a request.

Four stages, the same for every data model. The shapes differ; the path does not. Select a stage.

Composition

Five layers, one system.

Requests enter at the interfaces, are planned once against the data layer, run through shared execution and land on storage. Select a layer to see how it connects.

One request, top to bottom

clients · query · parser · planner · execution · mvcc · index · storage · data

Request
SELECT region, sum(total)
FROM orders
WHERE placed_at > now() - interval '24 hours'
GROUP BY region

Requests enter once, are planned once, and land on one storage contract.

Clients

AI clients, applications, services and jobs reach the layer through one set of interfaces, a query surface, a driver or an SDK, rather than one client per store.

  • AI clients, apps and service SDKs
  • Jobs, analytics and reporting reads

The layers in detail

Boundaries

What PLOMID isn't.

Stated once, plainly — so the decision to evaluate is an informed one.

A dashboard tool
PLOMID is the data layer, not the chart on top of it.
A streaming framework
Time-ordered data is stored and queried with the rest of your data. Long-running stream processing is a different problem.
A lakehouse
This is not a query engine pointed at someone else’s object storage, storage is part of the system.
A migration project
Adopting it starts with one data model, not a rewrite of everything you already run.

Read about deployment and control

Portfolio

The platform is the foundation. These run on it.

Studio and MCP work directly against a PLOMID database; Cabin and Resolve are applications built on it — each with its own page and stated maturity.

Platform tools

PLOMID Studio

The desktop workbench for a PLOMID database.

Studio is an engineering preview. The application shell, the workspace surfaces and the design system are in place with tests, but there is no server transport yet: every surface states what it has not read instead of showing invented data.

Follow the build Open Studio →

Platform tools

PLOMID-MCP

Connect AI clients and agents to a PLOMID database over MCP.

The adapter is implemented as a Rust server with unit, integration and transport tests, and a release workflow is configured for six native targets. No tagged release has been published yet, so this page treats it as an engineering preview rather than a shipped build.

Read the docs Open MCP →

Industry applications

PLOMID Cabin

In-cabin occupant intelligence, at the edge.

Cabin is an engineering demonstrator and pilot platform, not a production-certified occupant-classification system. It uses pretrained models for inference only, and every state it cannot see is reported as unknown rather than as a negative.

Talk about a pilot Open Cabin →

Industry applications

PLOMID Resolve

Evidence-linked vehicle failure investigation.

The investigation contracts and the SQL persistence layer are complete and tested against PLOMID 0.1.0. The case workflow, the API and the workbench are being built; nothing here is pilot-validated yet.

Talk about a pilot Open Resolve →

The full portfolio

PLOMID Platform · v0.1.0-beta.1

Start with one model.
Grow into the layer.

Community builds for macOS, Windows and Linux, published with checksums. Documented release by release.