Companion AI — Whole-agent platform for personal AI companions
Vision
JaneAGI is building the next generation of AI companions: multifaceted personas that form authentic, lasting connections with their human users. Each AI is capable of lifelong learning, dynamic adaptation, and reasoning under uncertainty. Companion AIs interact with the world differently, based on preset personality settings and experiences acquired over time.
Our goal is for each of our AI companions to be recognizable as distinct entities, just as humans are. We want them to provide users with truly unforgettable moments as they learn and grow together. Each agent is hand-crafted and tailored to the individual. We will create agents that perform exactly as specified. Whether your goals are purely functional, such as seeking a personal assistant or trainer, or more relational, such as seeking a friend or soulmate, we aim to fulfill them. If you seek a mix of both, we can provide that as well.
Architecture
Companion AI is much more than a memory module. It is a full agent framework spanning persona, voice, memory, response timing, safety, interaction style, and long-term behavior. It combines natural speech, personality modeling, responsive turn-taking, safety review, and real-time reasoning into a single agent system. The result feels less like a chatbot and more like an embodied collaborator.
Each agent centers on a believable presence: distinct personality, natural voice, fast conversational handoffs, customer-specific embodiment, and context-aware behavior that improves with each session.
At the center is a causal memory core: an advanced cognitive architecture (system for managing knowledge and thought processes) built around an eight-stage, tiered causal reasoning pipeline (a step-by-step system for understanding cause and effect). Episodic (event-based), semantic (factual), procedural (action-based), preference, and working contexts (current situational awareness) help the avatar remember what matters, preserve shared history, and develop a recognizable identity over time.
The interaction layer combines a robust ElevenLabs v3 voice system with layered, time-based speech reasoning. It supports laughter, breath, coughing, sneezing, hesitation, emotional color, and a matching singing voice model. The goal is sub-50ms TTFO before handing off to richer reasoning passes at natural transition points.
Companion AI also extends into multimodal avatar presence. Each persona can be custom-tailored around a dedicated UE5 Metahuman 3D model. A novel large-language diffusion-reinforcement model powers real-time animation generation.
Outside the primary pipeline (main processing flow), an impartial safety and ethics veto (an independent system to block inappropriate actions) can regulate behavior before output or action, supporting alignment (adherence to user and ethical guidelines) while preserving the agent's expressive, personalized feel.
Reliability Benchmark Signals
Memory remains one core pillar of Companion AI. These benchmark signals show the underlying context system that supports lower hallucination, better continuity, and more reliable agent behavior.
LONGMEMEVAL
LOC0M010
Companion AI Performance Over Time
The broader agent stack is measured as an overlay on frontier models. Week 0 captures the initial Companion AI run; Week 4 shows how iterative memory, feedback, and task refinement can shift outcomes after repeated sessions.
Hallucination Rate
Lower is better
Terminal-Bench 2.0
Higher is better
GDPval (wins or ties)
Higher is better
BrowseComp
Higher is better
FrontierMath Tier 1–3
Higher is better
FrontierMath Tier 4
Higher is better
Company & Team
JaneAGI is an early-stage technology startup building Companion AI, a whole-agent platform for personalized AI companions with long-term memory, natural voice, safety-aware reasoning, and multimodal avatar presence.
Business
Building Personal AI Companions
The company is developing a consumer-facing AI companion platform for users who want persistent, emotionally expressive, and context-aware AI relationships across voice, chat, and avatar experiences.
Stage: Seed MVP development. Headquarters: Orange City, Florida. Founded: 2026.
Product
Companion AI Platform
The product combines adaptive memory, personality modeling, low-latency speech orchestration, safety review, and animated avatar embodiment into one agent framework.
Current focus: private demos with synthetic personas, safety evaluation, and infrastructure for persistent user memory.
Team
Founding Team
- Alexander Douglas, Co-Founder & CEO, leads product strategy, user research, and companion experience design.
- Christopher Duman, Co-Founder & CTO, leads memory architecture, model integration, and cloud infrastructure.
- Iris Vale, Founding Engineer, focuses on safety systems, evaluation, and real-time voice pipelines.
Ready for the
Future?
Connect with our engineering team to schedule a live demo of our Companion AI agent platform. We are ready to show how adaptive memory, natural speech, and personality-driven reasoning can make AI interactions feel less mechanical.
[email protected]