The Assistant App

The Assistant is Delos's AI conversation app. It's not a simple chatbot - it's an agent capable of taking multi-step actions autonomously: writing documents, conducting research, building web apps, managing tasks, and connecting to your third-party tools (Slack, GitHub, Notion, Linear, Stripe, and more).

Video overview

Assistant Delos — Your AI Companion at Work

The Delos Assistant is much more than a simple chatbot: it's a conversational AI copilot that understands your needs, adapts to your context, and automates your daily tasks directly in the chat.

Key Features of the Assistant

1. Advanced Conversational Responses

  • Natural dialogue, context understanding, intelligent reformulation.
  • Structured responses in markdown, tables, bullet lists, code blocks, quotes.
  • Follow-up question suggestions to deepen or pivot.

2. Web Search & Sources

  • Real-time information search on the internet.
  • Reading and synthesis of complete web pages.
  • Automatic source citations (URL, titles, excerpts) in messages.

3. Location & Maps

  • Search for addresses, businesses, places, restaurants via Google Places.
  • Display results on interactive maps integrated into the conversation.

4. Image, Product and Video Search

  • Search for images on the web to illustrate your exchanges.
  • Search for products (Amazon, Google Shopping) and relevant videos.
  • Results displayed directly in the conversation.

5. Advanced Calculations & Mathematics

  • Solving mathematical problems, calculations, statistics, result visualization.

6. Email Drafting

  • Generation of professional email drafts, response suggestions, thread tracking.
  • Preview and editing before sending, directly in the chat.

7. Task & Plan Management

  • Creation of task lists, action plans, visual roadmaps.
  • Tracking and organizing priorities in the conversation.

8. Custom External Tools

  • Addition of custom APIs by your organization (URL, API key, parameters).
  • Configuration in Memory/Tools settings for specific uses.

Maestro — Intelligent AI Action Orchestration

Maestro is the orchestration engine that allows the Assistant to chain multiple actions autonomously and coherently. It automatically breaks down your complex requests into successive steps, calls the necessary tools, and manages the logical flow between actions.

Example:

"Summarize a web page, then propose an email to share the summary." Maestro coordinates each step and delivers the final result without manual intervention.

Benefits:

  • Time savings
  • Workflow fluidity and automation

Unlimited Mode — Unleash the Power of the Assistant

Unlimited mode allows you to use the Assistant without quota or credit constraints on AI requests. This mode is reserved for users with a specific subscription or enterprise access.

How it works:

  • All features are accessible without volume or action number restrictions.
  • Perfect for intensive use: brainstorming, massive writing, repeated analyses, automations.

Activation:

  • Unlimited mode is activated in your account settings or via organization subscription.
  • The interface clearly indicates if you are in unlimited mode.

Voice Dictation — Generate Prompts by Voice

Voice dictation allows you to enter your requests to the Assistant by speaking, without having to type your prompt. The Assistant automatically converts your voice to text, then executes the request like a classic prompt.

How it works:

  • Click the microphone icon in the Assistant input area.
  • Speak your request clearly ("Summarize this document", "Create a summary on the 2024 budget", etc.)
  • The dictation is transcribed to text and sent to the Assistant.
  • You can correct or complete the text before validation.

Benefits:

  • Speed: formulate long or complex requests effortlessly.
  • Accessibility: ideal for those who prefer voice or work on the go.
  • Precision: transcription is optimized for structured prompts.

Intelligent Memory System

The Delos Assistant retains and leverages your context to personalize each interaction:

  • User Profile: name, position, email, communication preferences.
  • Company Profile: name, sector, description, reference documents.
  • Preferences: tone, style, favorite formats, response length, emoji usage.
  • Memory elements: facts, projects, contacts, workflows memorized on request.

Total control: Memory can be enabled or disabled at any time, conversation by conversation.

Artifacts Panel

  • Artifacts generated by the Assistant (emails, tasks, plans, search results, etc.) are accessible in a dedicated panel to the right of the chat.
  • Visualization, editing, and centralized tracking of artifacts linked to each conversation.

Anatomy of a Message

Each message can contain:

  • Formatted text (markdown)
  • Visible reasoning (reflection mode)
  • Operation status (search, editing, creation)
  • Cited web sources
  • Location maps
  • Email and event previews
  • Task plans
  • Follow-up question suggestions
  • Generated or found images

Real-Time Streaming

  • Responses streamed word by word, with activity indication ("Searching...").
  • Progressive display of artifacts, sources, images, or other results.
  • Ability to stop generation at any time.

Daily Kickoff & Home Screen

  • Start your day with an automated summary: unread emails, upcoming tasks, events, updates.
  • Home screen: quick input bar, access to recent conversations, "Autopilot" quick start, filterable conversation gallery, bulk actions.

Understanding Limitations

  • Context window: Long conversations are automatically summarized to ensure performance.
  • Memory: Must be enabled to benefit from personalized context.
  • Internet access: Necessary for real-time searches (AI models have a cutoff date for training data).
  • Artifacts: Linked to each conversation, accessible at any time.
  • External tools: Some custom applications require prior connection by your organization.

Ready to experience an AI that truly works for you?

Start a conversation and discover everything the Delos Assistant can accomplish.

AI reasoning level

Before sending a message, a selector lets you choose the AI reasoning level: low, medium, or high.

  • Low: fast response, suited for simple questions or rephrasing tasks.
  • Medium: balance between speed and depth of analysis.
  • High: deeper reasoning, recommended for complex problems and multi-criteria analysis. This level consumes more credits.

The selected level applies to the current message only.

Source: app.delos.so, PR merged August 28, 2026.