Inference
Depth, memory, and job execution on the local network — context from your systems, not a public service.
The on-site mind
Aella
On-premises site intelligence
On-site intelligence for inference, memory, jobs, and workflows — installed on your infrastructure.
Aella runs on the site’s AI servers. Operators access it from the Stella workstation console and from Skjal on mobile devices, including existing conversations and groups.
Aella is the on-site mind. Stella is the operator console. Skjal is human chat; Aella-in-Skjal is a paid add-on and is plaintext to her.
Access
Workstation access through Stella. Mobile access through Skjal. Direct messages and existing groups.
Operations
Jobs, workflows, and approvals on site. APIs and agents for work that should not wait on a conversation.
How operators reach Aella
The Aella installation includes the office workstation and mobile access. One product. Two operator surfaces.
Stella
Operator console for Aella
The operator console for Aella. Conversations, jobs, status, and media. Not every desk — the ops surface. Opened from Saga with the same Rune session.
Skjal
Mobile site communications
End-to-end mobile communications between people. Aella can join a room as a paid add-on; that path is decrypted so she can reply, and is disclosed. Human-to-human rooms stay end-to-end.
Customer sites have no public widget. This site’s demo chat is schedule-limited and talks to our reference Aella at Strangstad Tech — not your production instance. Models and memory on a customer site stay on that site unless you connect an external service.