v0.2.0

As we run more pilots, we're learning what works and what doesn't. The core principle of Digel is that it must be easy to use and easy to adapt. There are many ways to describe a process, and more than one can be correct. The only way to learn is by doing.

Canvas

The process canvas is where you draw and map your processes (think Miro, but for complex industrial processes). We've made it easier to create nodes, relationships, and add descriptions and tags.

Digel AI

Improved semantic search helps the agent find relevant data faster, even across complex and larger processes.

Code execution

The agent can now write and execute Python code in sandboxed environments, enabling more complex data analysis with raw timeseries data.

Dashboards

Ask Digel AI to create a dashboard and it will generate a live, interactive Streamlit app for you. Share it with your team via a simple link.

File uploads

Upload files directly to the chat and the agent will parse and analyze them for you. Supports CSV, Excel, JSON, and more.

Data & Connectivity

Two major updates for getting data into Digel:

Integrations Orchestration

Using Kestra as the orchestration engine, you can now write flows to fetch data from external systems and send it to Digel. Learn more about Kestra here.

Timeseries API endpoint

Send timeseries data directly to Digel via a simple API call. Any system that can make HTTP requests can now push data to Digel.

curl -X POST 'https://<your-tenant>.digel.io/api/telemetry/ingest' \
  -H 'apikey: <your-api-key>' \
  -H 'Content-Type: application/json' \
  -d '[{ "timestamp": 1702900000000, "source": "sensor_1", "tag": "temperature", "value": 23.5 }]'