Camlin Connect
Dated releases across the Camlin Connect family, taken from the platform changelog. Everything listed is live today. Last updated 23 Oct 2026.
Connect
Accounts, Create, Companion, Team, Architect, speech recognition and AgentLink run in AWS US East (N. Virginia) for US customers, with US phone numbers and registered US texting. nuamedia.com is the United States & World site; support comes from our team in Australia.
Companion
Create Plus and Pro (A$35 and A$89 a month) and Companion Front Desk (A$495 founding) and Front Desk + Phone (A$995) can be bought online. Australian prices include GST.
Team
Camlin Team opens its own portal: rooms where specialist avatar seats, built in or played by your own agents over MCP, talk to you and to each other, with a live action plan and documents. Enterprise preview.
Create
Create reaches full release with physics-based 3D Studio, talking avatars, Camlin Drafter documents, executable research papers, a 1 GB asset library, camlin.data and live co-editing.
Companion
Create can add a Companion to a hosted website, while operators can recrawl its approved knowledge from Camlin Ops.
Connect
Camlin Server adds functions, shared collections, Architect Server nodes and dedicated single-tenant instances for customer workloads.
Create
Create websites can publish to a dedicated <slug>.create.camlinconnect.com.au address with the project kept behind the same governed release path.
Create
Anyone can make an account and talk to a Camlin avatar that makes games, pages, 3D models and Camlin Connect assistants on a canvas.
Companion
One immutable release now serves every Camlin site. It adds the avatar loop runtime, a reflex overlay and speech settings per avatar.
Avatar
Auto, HD or SD rendering per surface, mood reactions driven by live sentiment, and a speech engine and phrase list chosen per avatar.
Companion
Site owners pick the phone number their Companion answers. A text to that number gets one grounded reply from the same knowledge pack.
Connect
Intent matching from authored phrases and patterns, with no model call. It runs under the No-AI runtime mode used in sovereign deployments.
VoiceID
Verifications, enrolments and voiceprint health in one operations view. Presentation-attack detection at enrolment, hosted in Sydney or on your own infrastructure.
Architect
Test apps are authored as interactions, so the canvas shows the real flow. New sentiment, mood and gesture test apps sit in the library.
Insight
Specialist agents can propose key revocation, release cut and release activation. Nothing destructive runs until an operator confirms it.
Speech
Self-serve plans on the Speech portal, a live continuous captioning demo, and Australian lexicon biasing for names, places and reference numbers.