If ChatGPT is an AI chatbot, Intueo is an AI workforce.
That distinction matters. After spending the past year building Intueo, we have come to see the assistant-versus-workforce distinction as the most important architectural choice in consumer and prosumer AI today. Most platforms have chosen the assistant path: one model, one personality, one shared memory, designed to be helpful with anything. Intueo took the other road. We built a studio where you assemble a team of specialized AI agents, each with its own identity, knowledge, memory, and tools, and switch between them as naturally as you would switch between coworkers.
This post is a tour of what we have actually shipped, why the architecture looks the way it does, and what makes Sage 1.5, the engine underneath, different from a context window with extra steps.
The Core Idea: A Studio, Not an Assistant
Open Intueo and you do not see a single chat. You see a roster of agents. One might be a nutrition and wellness coach. Another might be a business analyst trained on your company's data. A third might be a creative writing partner that knows your voice. A fourth might be an automation specialist with deep tool integrations. You can have as many as you want, each tuned for a specific kind of work.
Every agent has four things that make it distinctly its own:
- A persona and system prompt that defines how it speaks, what it values, and how it approaches problems
- A persistent memory that survives across sessions, channels, and weeks of inactivity
- A knowledge base of documents, PDFs, and notes specific to that agent's domain
- A toolset that determines what it can actually do, from web search to custom MCP-connected tools
Switching between agents is one click. Each agent is isolated, so what you tell your nutrition coach does not leak into your business analyst's context, and vice versa. This is closer to managing a team than chatting with a bot.
Sage 1.5: Three Modes, One Brain
The model layer powering Intueo is Sage 1.5, our proprietary orchestration and memory layer. Sage 1.5 ships with three execution modes, exposed to users as a simple toggle:
- Sage 1.5 Fast — Quick responses, vision-capable, optimized for everyday messages and image attachments
- Sage 1.5 Thinking — Balanced reasoning for moderately complex problems
- Sage 1.5 Deep — Heavy reasoning for genuinely hard problems where you want the model to think carefully
Users pick the mode that fits the task. But here is one of the small details that we are most proud of: when you attach an image to a message, Intueo silently switches to Fast mode for that single message because Fast is currently the only vision-capable variant, sends the message, and immediately switches back to whatever mode you had selected. Your toggle stays exactly where you put it. You never see the swap. The platform handles the technical reality so you do not have to.
This kind of invisible engineering shows up everywhere in Intueo. It is also the philosophy that has guided our entire build: hide the complexity, expose the capability.
What Makes Sage 1.5 Different
To understand why Sage 1.5 is not just "another wrapper," it helps to compare it directly to what users get from generic AI products today.
Real Memory, Not Just Context Window
ChatGPT's memory is a feature layered on top of a fundamentally stateless model. Claude treats every conversation as its own world. Gemini has its own approach but is primarily session-bound.
Sage 1.5 uses a different architecture entirely. Every agent has core memory (always loaded), archival memory (long-term searchable storage), and recall memory (past conversations). These persist between sessions automatically. If you told your wellness coach a month ago that you are vegan, it remembers a month later. If you told your business agent to always respond in formal English, it keeps doing that forever, across every channel.
This is not an extended context window. It is genuine persistent state, and it changes how the agent feels to use.
Multiple Specialized Agents Instead of One Generalist
ChatGPT tries to be everything to everyone. The result is an assistant that is decent at almost any task but specialized for none.
Intueo flips that. Instead of one generalist, you get a team of specialists. Each agent gets dedicated training (memory, prompt, knowledge), a separate persona, and a focused toolset. You build the team you actually need rather than relying on one model to context-switch between every domain in your life.
Agent Workshop: No-Code Agent Building
Custom GPTs and Project Claude have moved in this direction, but they are constrained. Intueo's Agent Workshop lets you:
- Edit the persona and system prompt directly
- Manually edit core memory (the things the agent always "knows")
- Upload knowledge files for retrieval
- Wire up custom tools, including MCP-connected services
- Configure reasoning settings per agent
- Cycle the model mode (Fast / Thinking / Deep) per agent
Every change applies live to a real agent you can talk to in the same moment. There is no separate "build" step, no deploy delay. You configure, test, and ship inside the same window.
Real Voice Calls With Persistent Memory
Voice in Intueo is not text-to-speech bolted onto a chat. It is full conversational voice with the same agent, the same memory, the same personality. You can call your wellness coach the way you would call a friend and talk through your weekly meal plan. When you go back to text the next day, the agent remembers the voice conversation. The interface is different but the agent is the same.
Most "AI voice mode" features are essentially separate products that happen to share a brand. Sage 1.5 voice is just another channel into the same persistent agent.
Genuine Multi-Modal Support
- Drag and drop up to ten images at once into a single message
- Silent vision-aware model swapping so you never think about which mode supports images
- Knowledge file ingestion for PDF, Markdown, and text
- Rich content rendering with markdown, code blocks, tables, and automatic RTL/LTR detection
- Streaming responses with a typewriter effect for natural pacing
- Expandable reasoning steps so you can see exactly how the agent thought
Admin Panel and Tools Manager
Intueo is a multi-user platform from the foundation up. The Admin Panel manages model availability, usage tracking, and team configuration. The Tools Manager lets users enable or disable MCP tools per agent. These are typically enterprise-tier features in other platforms. In Intueo they are the default architecture.
PWA, Mobile-First
Intueo is a Progressive Web App. You can install it on iPhone or Android, launch it offline, open it from a home-screen icon, and use proper iOS keyboard handling. Most AI platforms are either web-only or have a separate native app that does not sync cleanly. Intueo is one platform that runs everywhere.
The Details That Make It Production-Grade
Some of what we are most proud of are not headline features. They are the engineering details that make the platform feel solid:
- Reasoning toggle that overrides per-message rather than per-agent, so you can dial in Deep reasoning for one hard question without changing the agent's default
- Self-healing chunk-load recovery so a deploy mid-session no longer crashes the app
- Per-message error boundaries so one broken message cannot bring down the whole interface
- Optimistic UI with proper fallback for assistant replies, so the interface stays responsive while memory indexes catch up
- Multi-image safe pipeline that handles batches without the duplicate-render bug we shipped and then fixed in week three
- Eventually-consistent message reconciliation between streamed responses and the canonical message store
These are not the features anyone puts on a marketing page. They are the difference between "demos well" and "you can actually live in it every day."
Why It Matters
ChatGPT is a universal assistant trying to do everything for everyone.
Claude is a strong-reasoning assistant that is fundamentally stateless.
Gemini is Google's product, deeply integrated into the Google ecosystem.
Intueo is a different category. It is a platform for building your own personal AI team, where each agent has lasting memory, a distinct identity, and the ability to do real work. The interface is simple enough for casual users (toggle, drag and drop, voice call) and deep enough for power users (agent workshop, tools, admin panel).
We built Sage 1.5 because we wanted something that ChatGPT could not be: not one assistant pretending to be everything, but a workforce of specialists that actually remember you, your work, and your context.
The platform is in active development and rolling out to early users now. If you want to see it in action, the best way is to get in touch for a walkthrough, or read more about the underlying memory system on our Sage 1.5 page.
