The LabMission platform

The platform under every LabMission application

Connections, data streams, queries, dashboards, agents and your own code, governed by provenance and access rights and reachable through APIs and an MCP server. Your team works on it on a visual canvas, an agent works on it through MCP, and both go through the same review.

LabMission/Biobank · Site A
Surface · Cold chain
Connection · MQTT
Freezer B · ULT sensor
-79.6°C
Connection · HTTP
Plate reader PR-2
Results file · column map v3
API source · every 15 min
Courier tracking
Expected arrivals
8 inbound
Warm query
excursions_24h
Published as an API
Agent
Excursion triage
Drafts impact items
Interface
Storage dashboard
Capacity · alerts
LIVE STREAM
14:32:20 freezer-b { temp_c: -79.6, door: closed }
14:32:19 pr-2 { plate: P-0412, wells: 96, status: read }
Why it matters

Applications inherit rather than reimplement

Attribution, audit, access control, retention and change control belong to the platform. Every application, every agent and every workbench inherits them, so your quality team learns them once.

  • Compliance behaviour belongs to the platform, so every application behaves the same way
  • One data model underneath, so information moves between applications without a project
  • New capability arrives for every application at once
  • What your quality team learns about one application holds for the next
Built and run by
  • People
    The canvas and the applications
  • AI agents
    The MCP server
  • Code
    The API, SDK and workbenches

All three work through the same proposals, rights and record.

The platform under every application
Connect
  • Connections
  • API sources
  • Simulators
  • Files & manifests
Data
  • Live streams
  • Queries
  • Data stores
  • History & export
Build
  • Surfaces
  • Interfaces
  • Agents
  • Workbenches
Govern
  • Provenance
  • Access rights
  • E-signatures
  • Change control
REST API & OpenAPI specMCP serverTypeScript SDKAPI keys
Your cloud
A workspace per tenantSecrets in your key vaultImmutable retentionData-residency policy
Building blocks

What a LabMission workspace is made of

Each block is a node on the workspace canvas. People place them by hand, agents propose them, and code deploys them, and all three are reviewed the same way.

Provenance

How every change becomes evidence

Provenance is configured per workspace: the hashing algorithm, whether approval and a second person are required, whether the before-and-after state is captured, and how long records are kept.

  1. 01
    A change
    By a person, an agent's approved proposal, or a process
  2. 02
    Its reason
    From the reason categories you define
  3. 03
    Its signature
    Review, verification, approval, authorship…
  4. 04
    Hashed & chained
    SHA-256, 384 or 512, linked to the last
  5. 05
    Retained
    Immutable storage, for up to 400 years
Access rights

Rights, roles and people, in three layers

Access rights say what can be done. Configurations bundle rights into roles. Access cards give roles to people and teams. The same layers govern people, agents and workbenches.

1
Rights
Run or draft a query, read or change a table, write files, deploy.
2
Configurations
Rights bundled into roles, such as a workspace administrator.
3
People and teams
Access cards that give a configuration to a person, a team or an agent.
APIs

Everything on the canvas is also an API

Workspaces, access, data stores, files, queries, live streams, history, provenance and licensing are all reachable over a REST API, described by a generated OpenAPI spec.

REST API
One API for the workspace and its data.
OpenAPI spec
Generated, and filtered to what the caller may use.
TypeScript SDK
A typed client for your own services.
API keys
Issued per workspace, and revocable.
MCP server
The same workspace, as tools for agents.
Query APIs
Any named query, published as an endpoint.
Query · freezer_excursions_24hPublished API
Readings
| where Timestamp > ago(24h)
| where Unit == "Freezer B" and TempC > -70
| summarize Minutes = count() by bin(Timestamp, 1h)
Where it runs

Three ways to run LabMission

Evaluation

A workspace provisioned for you, to try with your own data.

Fully managed

We run and operate the platform so your team can focus on the science.

Your own cloud

Deployed into your own Azure subscription through a landing zone: key vault, logging, monitoring and a data-residency policy.

See the platform underneath a real application

We'll walk a live workflow in Sample Management and open the workspace behind it: the connection, the query, the proposal and the record.