PlatformPrototype to production
Convai
Voice-first characters with backstories, knowledge, actions, memory, emotions, vision, and Unity, Unreal, web, and XR integrations.
Research signalAction and animation hooks
Official source ↗Recovered comparison / research snapshot
Eight signals from the embodied-agent market.
This is a readable preservation of the Character Current directory compiled 14 September 2026. It describes documented positioning—not hands-on testing, ranking, endorsement, or current procurement advice.
How to read this page
Product availability, prices, rights, and integrations can change. Open the official source and verify the current offer before making a decision. “Strongest signal” means the clearest differentiator in the recovered research, not a measured winner.
PlatformVoice infrastructure
Inworld
Realtime AI infrastructure spanning speech synthesis, recognition, model routing, and character-oriented experiences.
Research signalLow-latency voice economics
Official source ↗PlatformHigh control
NVIDIA ACE
Modular digital-human technologies for speech, intelligence, vision, translation, animation, and NIM-based deployment.
Research signalModular deployment
Official source ↗Character stackCinematic fidelity
MetaHuman + Unreal
High-fidelity rigged humans and realtime animation; a separate conversational system is still required.
Research signalCharacter quality
Official source ↗Open sourceLocal-first
Utsuwa
A VRM-based companion reference combining avatars, voice, memory, and user-supplied models.
Research signalOwnership and portability
Official source ↗Open sourceCustomizable
Amica
A 3D avatar companion frontend with voice chat, vision, emotion behavior, and API-driven setup.
Research signalHackable embodiment
Official source ↗Consumer appiOS native
Companion AI
An animated anime companion described by its maker as combining voice, memory, and a local VRM-based scene.
Research signalPersistent presence
Official source ↗Hardware / appEvolving
Razer AVA
A holographic or desktop companion concept positioned around memory, mood, and audio-visual sensing.
Research signalAmbient presence
Official source ↗Planning ranges retained from the recovered edition
Budget by ambition, not by avatar alone.
These are editorial planning estimates, not quotes. Character labor, integration, inference, rendering, moderation, and support can move totals by an order of magnitude.
| Builder | Illustrative stack | Planning estimate | Proof target |
|---|---|---|---|
| Solo creator | VRM + Blender + hosted voice + simple web or desktop runtime | $0–$300/month plus time | Prototype the interaction loop |
| Indie developer | Unity or Three.js + a character platform + custom persona and memory | $300–$2,000/month | Ship one character to one audience |
| Small startup | Web or Unity client + modular realtime voice + owned orchestration and analytics | $2,000–$15,000/month | Validate retention, conversion, or task completion |
| Enterprise | Private or vendor platform + SSO, data boundaries, SLA, moderation, and handoff | $25,000+/month | Support, training, hospitality, or regulated workflows |
| High-end digital human | MetaHuman + Unreal + capture + custom animation and speech pipeline | $100,000+ build | Premium brand, film, spatial, or event experiences |