Streaming chat & tools
Tab، Cmd+K، Composer، chat، agent — jab plan، keys، uptime allow karein۔
Models
Online aur offline — ek registry، AUTO routing۔
Master registry: 206 online metadata، 49 offline GGUF۔ Modes: ONLINE، OFFLINE، AUTO، MANUAL۔ Ready wizard AUTO (online-first) default۔ Offline: ORQVION AI + GGUF (llama.cpp) — Ollama app nahi۔
Online
Entitlements server par؛ inference desktop par۔
License server control plane — plans، seats، quotas؛ prompt proxy nahi۔ Desktop → provider APIs (BYOK keychain) jab online ho۔
206 online catalog
Metadata routing/entitlement — har id hosted endpoint nahi۔
Honest online limits
BYOK، provider health، plan۔ OpenAI required nahi۔ Org policy external models rok sakti hai۔
AUTO routing
Surface aur network ke mutabiq model۔
Task classifier + hardware/network۔ Manual pick fateh۔ Fallback: alternates → local GGUF → ORQVION AI bundle۔
Tab completion
Low-latency model — routing mode ke mutabiq۔
Agent & Composer
Long context — online/BYOK ya local GGUF fallback۔
Context budget
@-mentions، rules، index — visible window۔
Offline & BYOK
GGUF machine par — cloud keys optional۔
ORQVION AI + HTTPS downloads (`.part` pause/resume)۔ BYOK: keychain — org policy gate۔
ORQVION AI bundles
Signed offline — llama.cpp locally۔
49 offline GGUF catalog
Model manager downloads — no ollama serve / :11434۔
Advanced Local
Settings → Advanced Local؛ Ready wizard skip — same binary۔