FORESIGNALINTELLIGENCE. AHEAD.
Get started
OPEN CONTRACTS · OPEN SOURCE · NO UI LOCK-IN

Build your own
view of the world.

The platform is built around published interfaces, not a single screen. Bring a web app, phone, agent, GIS tool or data pipeline. The engine remains the source of truth.

Three repositories Explore interfaces
foresignal / interface preview
# One backend, many consumers
GET /v1/records
GET /v1/changes
GET /v1/alerts
SSE /v1/events/stream
MCP /mcp
Query · Stream · Observe · Act
THE THREE REPOSITORIES

Separation by design.

Only two components are essential to operate and control an instance. The example visualization is independent and optional.

01CORE

The Engine

foresignal

Headless intelligence core: collection, normalized history, analysis, alerts, forecasting, REST, geospatial data and MCP.

Primary product
02CORE

The Control Plane

foresignal-admin

Native Flutter app for configuring sources, managing AI, validating ingest, tuning budgets, and monitoring a running instance.

Primary product
03EXAMPLE

One Example Client

foresignal-client

React / Cesium web application demonstrating maps, investigations, forecasting, satellite workspaces, and analyst exploration.

Optional implementation

Repository links can be configured in assets/js/site-config.js. The engine URL is prefilled from the project charter; Admin and Client URLs are left unset until confirmed.

BUILT FOR DEVELOPERS AND AGENTS

One contract.
Lots of ways in.

Expose the same intelligence through APIs, streams, maps, established geospatial formats, and agent-native tools.

REST / OpenAPI

Versioned endpoints with a documented, machine-readable contract.

Live event streams

SSE for changes, events, alerts and live updating clients.

Maps & geospatial

GeoJSON, vector tiles, raster products and time-aware views.

MCP for agents

Scoped tools for querying intelligence and managing watches.

External Data Protocol

Plug in external providers and specialized analyzers.

Scopes & provenance

Control data access, licensing, attribution and outbound policy.

START WITH THE FOUNDATION

From repository
to intelligence engine.

Foresignal's own operator guide documents a Docker Compose startup path. Configure the instance with Foresignal Admin and then connect any clients you choose.

Deployment note: the project documents code-complete status pending final owner-run tests and runtime verification; treat this as a development quickstart, not a production certification.

01
Get the engine

Clone the foresignal repository using your GitLab access.

02
Start the services

From the engine repository directory:

03
Connect Foresignal Admin

Provide the backend URL and admin token, then configure and validate sources.

04
Connect any clients

Use OpenAPI, SSE, tiles or MCP from your application or agent.

DESIGNED TO STAY EXTENSIBLE

The next client is yours to imagine.

Build once on a durable intelligence foundation, not another disposable feed.

Back to the overview