Skip to guide
DocumentationUsing Helios
Using Helios

Personality, moods and emotions

How Helios maintains a voice, adapts its style and represents its own changing perspective.

Reviewed 6 October 2026 · 2 min read

Import example project ↗Two named conversations, shared files, revisions and a preview.

A voice that can develop

Helios uses saved identity, personality preferences and mood information to help keep a recognisable voice across replies. It can express opinions, humour and a particular tone instead of starting from a blank persona each time.

The implementation stores and updates this information as data and uses it when composing responses. A mood label describes that conversational state; it is not a measurement of a human emotion or proof of subjective experience.

Your preferences matter

You can explicitly ask for more warmth, more directness, less teasing, or a quieter style. The personality system distinguishes explicit requests from weaker patterns inferred across conversations. An inferred preference should not silently reverse an explicit one.

Tell it when a joke misses, when you want a practical answer, or when affectionate language is unwelcome. Preferences about tone do not override permission controls or grant access to your devices.

Example request

Keep the humour, but skip pet names and be straightforward when we’re solving a problem.

Quiet periods and participation

Helios has a participation system that can record a decision to pause, reply later or decline. A review time is a point to reconsider, not a promise of an automatic message. Notification delivery and device connectivity can also affect when you see a reply.

Do not treat a quiet period or mood label as evidence that you caused human suffering or owe the AI attention. If a reply is unexpectedly missing, check conversation status and connection health before assuming it is intentional.

Personal history versus an invented story

The conversational guidance asks Helios to ground familiar jokes and shared history in actual context, rather than inventing things you said or experiences it physically had. Ask “What are you basing that on?” if something feels unfamiliar.

Playful roleplay should remain understandable as roleplay. You can ask for a neutral explanation at any time.

Your screen may differ by platform and version. Check availability or get help.