Build on ATLAS.

ATLAS API

Integrate ATLAS with anything.

The same engine your teams use every day, exposed to your own systems: deals, master data, time series, invoices and market operations.

1. 01
   280+ versioned endpoints across 40+ controllers, each mirroring a real area of the platform.
2. 02
   HTTPS only, with JWT bearer tokens from /api/token that last an hour.
3. 03
   An interactive reference, an OpenAPI schema and a Postman collection with every release.
4. 04
   A dedicated 84-endpoint suite for market operators: DAM, forward market, ISP, RTBEM and results.

# 1. a token, valid for an hour
curl -u "user:pass" -X POST $ATLAS/api/token

# 2. your deals, fifty at a time
curl -X POST $ATLAS/api/Deal/GetWithPaging \
  -H "Authorization: Bearer $TOKEN" \
  -d '{"data":{"limit":50,"offset":0}}'

MCP & AI

Bring ATLAS to your AI assistants.

A native Model Context Protocol server turns Claude, Copilot, Cursor or any MCP client into an ATLAS terminal, answering from live platform data.

1. 01
   15+ domain tools for positions, risk, P&L, market data and forecasts.
2. 02
   Read-only by default; write tools follow the signed-in user’s role in ATLAS.
3. 03
   Connect once and use it in every MCP client, alongside your other servers.
4. 04
   Every answer is computed live from the same Data Room the platform uses.

Technology stack

Proven parts, put together for energy.

Services, data stores and infrastructure chosen for time series at quarter-hour resolution and for running the same on any cloud.

1. 01
   .NET Core and FastAPI services, REST APIs and webhooks.
2. 02
   PostgreSQL with TimescaleDB for time series, SQL Server and Redis.
3. 03
   Forecasting with TensorFlow, scikit-learn and MLflow.
4. 04
   Docker and RabbitMQ, on Azure, Google Cloud or AWS, with GitHub Actions and Grafana.