Documentation

The agentic task environment: Taskspace, engines, previews, and governance.

Scheduled Rivine engine runs that continue whether or not your computer is on.

Tokens, cost, latency, and Artificial Experience activity for your own turns.

Organizations, roles, and personal connections across the platform.

Base, Pro, and Essential plans, credits, and feature comparison.

Use Aselius from any AI app: connect Claude, Cursor, ChatGPT or any MCP client to https://workspace.beunec.com/mcp.

Start and follow a Rivine engine task from Telegram, Discord or WhatsApp.

preview.aselius.beunec.com: a deliverable rendered exactly as generated, with access codes, theme and language.

Why Aselius is agentic infrastructure software, not another Claude Cowork, ChatGPT Work or Kimi Work.

Security model, data handling, encryption, and vulnerability reporting for Aselius.

Action-oriented professional work: connected apps, specialties, and deliverables, not generative chat.

Source-grounded investigation delivered as a report or a structured spreadsheet.

Plans, delegates, validates, and previews software in a governed sandbox.

Legal-domain engine with conditional delegation and honest gap escalation.

Seven media agents: generate and surgically edit images and video, record speech, compose music.

Country forecasting and scenarios: models choose evidence; code computes the numbers.

Surgical follow-up edits on a document already open in Preview, without a full engine turn.

Browser agent in your real tabs and sessions. Opens the Browse product paper.

Opens Rivine Browse →

Workspace memory of facts, outcomes, and reusable work so engines reuse what already succeeded.

Retrieval under a cap so engines operate over large corpora without sending millions of tokens per call.

Rivine Code in your terminal: local files, streaming commands, and permission modes.

Speech-to-speech voice for Aselius and Rivine Browse, with the ivory-and-jade dragon.