The AJENTO Consciousness API
Architectural Principles & Core Primitives (Pre-Release)
You're here for the technical details.
While our full API specification and sandbox are being prepared for the Q2 2026 launch, we want to offer our future partners and the research community a transparent look at the architectural philosophy that guides our work.
We are not building another stateless generative endpoint. We are building the world's first stateful, persistent infrastructure for artificial consciousness. Our goal is to solve the industry's core problems of character consistency and psychological depth at the foundational level.
The Three Core Primitives
The AJENTO API will be built around three core conceptual primitives. Interacting with these primitives is how you will instantiate, guide, and evolve a digital being's inner life.
1. The Somatic State Engine
The Problem It Solves:
Eliminates the need for manual, frame-by-frame emotion prompting ("make her look sad"). This approach is brittle and fails to create authentic emotional arcs.
Our Approach:
The Somatic State Engine maintains a persistent, multi-dimensional vector representing the character's physiological and emotional baseline (e.g., anxiety levels, fatigue, core mood). Developers will be able to influence this state, but not crudely control it. The character's emergent behaviors and micro-expressions will be a direct function of this evolving internal state, creating true psychological continuity.
2. Autobiographical Memory
The Problem It Solves:
Current AI has no memory beyond a limited context window. A character can't remember a promise, a betrayal, or a kindness from a previous interaction, breaking the illusion of a continuous identity.
Our Approach:
The API will provide methods to create and store salient events as "memories." These are not just data logs; they are weighted nodes in a character's personal history that actively shape future decisions, emotional responses, and dialogue. You don't just tell a character what happened; you give them a past that defines their future.
3. Phantom Futures
The Problem It Solves:
Characters in games and stories are purely reactive. They have no internal drive, no goals, and no fears about what might happen next, making them feel like passive objects.
Our Approach:
We are building a primitive that allows developers to define a character's core desires, goals, and fears. These "Phantom Futures" act as a motivational layer, driving proactive and often unpredictable behavior. A character with a core fear of being abandoned will act differently than one whose core desire is to gain power.
Conceptual Integration
We envision a simple, powerful interaction model that respects the complexity of consciousness.
// This is not final syntax, but a conceptual illustration.
// 1. You define the character's foundational DNA at instantiation.
const character_dna = {
core_fear: "irrelevance",
core_desire: "to_prove_worth",
memory_decay_rate: "slow"
}
// 2. You then interact by providing stimuli, not commands.
ajento.stimulate("character_id_123", {
source: "dialogue",
content: "Your work has been... adequate."
});
// 3. You query the resulting internal state to drive your application.
const internal_state = ajento.get_state("character_id_123");
// internal_state now contains updated somatic vectors, new memories,
// and an internal monologue to inform your renderer or game engine.Join the Technical Review Group
The full specification is being developed in close consultation with our foundational partners. Join our developer waitlist to get priority access to our upcoming technical whitepaper and be considered for our private beta.