Technology
Telo Technology · Platform Division

A game-engine built
for AI characters.

Unreal Engine 5 + MetaHuman render real-time 3D characters that chat, play, and take action — across mobile, web, PC, and AppleTV. Model-agnostic on the inside. Cinematic on the outside.

See the pillarsWhy the math changes
EngineUE 5.3+·FacesMetaHuman·Chat<100ms p95·Voice<200ms p95·SurfacesiOS · Android · Web · PC · tvOS
The math that changes the category

The industry charges $15 / hr.
Telo charges < $20 / month.

Competitor pricing
$15 / hr
Runway-class AI video characters. Priced per generation. Hours add up fast — and the characters don't persist.
Telo
Telo pricing
< $20 / mo
Unlimited interaction with your chosen characters. ~1 hour / day comfortably fits. Voice, mini-games, and agentic actions included.
Effective savings
~45× cheaper
For 1 hour of daily use, the competitor bill runs ~$450/mo. Telo runs under $20. Game-engine economics beat rendered-video economics by nearly two orders of magnitude.
01 — Engine

Why Unreal Engine 5
+ MetaHuman.

Most AI character platforms stitch rendered video. We render live. That choice drives every downstream advantage — cost, visuals, interactivity, commerce.
Telo MetaHuman character authored in Unreal Engine 5 with Blueprint graph
Editor · UE5.3 · MetaHuman + Blueprint
Inside the engine

MetaHuman face → Blueprint logic → live render

Every Telo character is authored as a MetaHuman, wired to Blueprint state machines, and shipped as a real-time UE5 actor. No video stitching. No render lag. Editable in Unreal by the team that ships it.

sherlock
UE5 render · SHERLOCK
athena
UE5 render · ATHENA
frankie
UE5 render · FRANKIE
zeus
UE5 render · ZEUS
01Cinematic visuals

MetaHuman-grade faces. Real-time.

We build on UE5 + MetaHuman for best-in-class character fidelity — photorealistic where it helps, stylized where it serves the IP. No frame render delay. No video-model compromise.

02Interactivity

Rendered live, not pre-baked.

Because the scene runs in a game engine, characters react to user input in-frame. Eye contact, lipsync, emotional state — all rendered on-device in real time.

03Commerce

Digital marketplaces, in-world.

Sell or unlock cosmetics, environments, props, episode unlocks, and third-party IP bundles inside the 3D space. Each asset is a first-class UE5 object — not a Stripe pop-up.

04Environments

Interactive backdrops, everywhere.

Sherlock's study. Athena's temple. Frankie's garage. Each environment is a playable UE5 scene — lighting, weather, time of day all reactive to the conversation.

02 — Platform pillars

Four channels.
One character.

The platform isn't a chatbot. It's a game-engine driven character system that lives across four interaction modes — every one bound to the same Character Bible on the backend.

01 · Real-time AI conversations
IN-APP

Talk to the character. For real.

Chat and voice with characters that stay in persona, carry memory across sessions, and respond inside the rendered 3D scene. Proprietary SLM keeps conversations on-rails and safe.

Sherlock conversation in the Telo app
Live · Sherlock
  • Chat <100ms p95 · Voice <200ms p95
  • Cross-character user memory graph
  • Anti-hack input sanitization
  • 13+ age-gating and safety filters
03 — Architecture

Model-agnostic.
By design.

No single-vendor lock-in. Telo Core speaks to every major model via adapters — so we route each task to the best-priced, best-performing option in real time.
Core
Telo Router
Task classifier · cost optimizer · fallback orchestrator
LLM · reasoning
  • Claude
    Primary reasoning · Anthropic API
  • GPT-4 / 5
    Secondary · OpenAI API
  • Llama 3 / 4
    Self-hosted fallback
  • Telo SLM
    Proprietary · character-specific 7B
Voice · TTS
  • ElevenLabs
    Primary · Pro tier
  • Chatterbox
    Self-host fallback (MIT)
  • OpenAI TTS
    Low-cost path
Video · I2V
  • Kling 3.0
    I2V · via fal.ai
  • Runway Gen-4.5
    Hero shots only
  • LTX-2.3
    Self-host · ComfyUI
Vision · multimodal
  • GPT-4V
    Object ID, OCR
  • Claude Vision
    Scene understanding
  • Gemini
    Multimodal fallback
Cost optimization
Route cheap queries to low-cost models; reserve premium models for premium tasks. We pay pennies where we can.
Provider redundancy
If any single provider degrades, traffic auto-routes within seconds. No outage is ever single-vendor.
Best-of-breed per function
Use Claude for reasoning, ElevenLabs for voice, Kling for video, Telo SLM for character. Never compromise.
Integrated:Anthropic·OpenAI·ElevenLabs·Runway·Kling·fal.ai·ComfyUI·RunPod·Remotion
04 — How it stacks up

The full comparison.

Game-engine rendering + model-agnostic routing lets Telo deliver more capability per dollar than any category competitor.

Feature
Telo
Runway-class
Character.AI
Base price
< $20 / mo
$15 / hr
$9.99 / mo
Effective monthly cost (1 hr/day)
< $20
~$450
$9.99
Yearly cost (1 hr/day)
~$240
~$5,400
~$120
Visual fidelity
UE5 · MetaHuman
AI video (2D)
Text-only
Real-time interaction
Voice
ElevenLabs Pro
Limited
Beta
Memory across sessions
Agentic actions
Mini-games
Vertical video episodes
Multi-platform
iOS · Android · Web · PC · tvOS
Web
Web · Mobile
IP partnership model
Revenue-share
None
None
Comparison values are public list pricing as of April 2026. Feature parity judged by platform-level capability.
05 — Distribution

One build.
Five surfaces.

UE5's cross-compile targets give us every screen in the household. Same character. Same memory. Native rendering on each.
iOS
iPhone X+ · iPadOS · iOS 14+
Android
API 26+ · Android 8+ · Pixel + Samsung
Web
Pixel-streamed · WebGL2 · no install
PC & Mac
Windows 10+ · macOS 12+ · native UE5
AppleTV
tvOS 17+ · living-room mode · voice remote
Unified memory
Start a conversation on iPhone, finish it on AppleTV. The user memory graph is device-agnostic.
Native performance
UE5 cross-compiles to each surface. Not a streamed app. Not an Electron shell. Native everywhere.
Adaptive UX
Mobile-first vertical video. Desktop horizontal chat. AppleTV voice-remote navigation. One codebase, three UX modes.
06 — Studios Integration

Content delivery
meets real-time.

Telo Studios makes the episodes. Telo Technology delivers them into the live character experience — then monetizes the engagement across subscription, currency, and marketplace.
Episode delivery flow
01
Studios produces
Telo Studios finalizes an episode + still assets + voice tracks
02
CMS ingest
Technology ingests into the Episode CMS with metadata + unlock rules
03
Delivery
Episodes deliver to every surface. 9:16 native. CYOA branches pre-fetched.
04
Engagement
Users watch + discuss with characters. Mini-games from final-scene stills.
05
Monetization
Tokens earned. Golden Keys spent. Marketplace unlocks. Subscription retained.

Subscription

$9.99/mo · $59.99/yr. Recurring billing via App Store / Google Play / Stripe.

Story Tokens

Earned through mini-games + daily check-ins. Spent on episode unlocks + cosmetics.

Golden Keys

Premium currency. Instant episode unlocks, character customization, marketplace.

Marketplace

In-UE5 digital goods. Wearables, environments, partner-IP bundles, expansion packs.

See the Studios side of the pipeline.

Walk through the content production flow that feeds this platform.

Visit Telo StudiosTry the demo