Data never leaves your premises.
All CALLAII models — ASR, NLU, sentiment, coaching generation — run on your own servers. Call recordings, transcripts and analytics never reach Callmenta servers.
CALLAII is a model, CALLAII Edge is a system.
CALLAII is the AI layer that listens and understands every call — it lives in the cloud and ships fast. CALLAII Edge brings that same intelligence to your premises, complete with hardware, audio stack, archive and license management.
CALLAII
AI model fine-tuned for Turkish. Runs on the Callmenta cloud.
- ✓ Fast deployment (5–10 business days)
- ✓ Software license / per-seat
- ✓ Automatic model updates
- ✓ Elastic scale, low ops load
- · Requires internet connectivity
CALLAII Edge
CALLAII model + ASR/TTS + archive + license portal + hardware — all on premises.
- ✓ Data never leaves the organisation
- ✓ Air-gapped (offline) capable
- ✓ Hardware + software + audio layer
- ✓ Annual on-prem license + update packages
- ✓ BDDK / KVKK / public-sector compliance ready
In short: CALLAII is the brain in the cloud; Edge brings that brain to your servers with the hardware and audio layer attached.
For anyone where data cannot leave the premises .
Banking & Finance
Customer interactions under BDDK regulation.
Healthcare
Patient data, KVKK and patient-privacy rules.
Public Sector
Citizen contact centers, classified operations.
Insurance
Claims records, sensitive customer information.
Legal
Attorney-client calls, case confidentiality.
Defense Industry
Classified communication, domestic tech mandate.
Every operation in one panel.
An application that runs on the Edge node: monitoring sources, matching flow, call analytics and license health — one window.
Sample panel view — actual data depends on your configuration.
3 deployment options.
Three deployment architectures keyed to your data-classification strategy.
Full Edge (Air-gapped)
No data leaves over the internet. Model updates apply through signed packages by hand or over intranet.
Edge + Management Cloud
AI models on Edge, only license management and metric summaries in the Callmenta panel. Call content stays local.
Hybrid
Sensitive calls on Edge, the rest on Cloud SaaS. Classification rules are yours, single pane.
3 packages for your scale.
Hardware, license and service level are set per package. Net pricing is tailored to your operations volume and sensitivity profile.
Pilot & Small Operation
Start with a single node; ideal for pilot validation or 50–150 seat operations.
- • 1 × GPU node (RTX 4090)
- • 50–150 concurrent seats
- • All CALLAII models local
- • Standard support (business hours)
- • Manual signed updates
Mid-scale + HA
High availability. Fully redundant architecture for 150–500 seat production operations.
- • HA pair (2 × A6000 / A100)
- • 150–500 concurrent seats
- • Auto failover, redundant storage
- • Priority support (8×5)
- • Edge + Management Cloud mode optional
Large-scale + Air-Gapped
Cluster architecture, fully offline operation — for critical infrastructure and regulated sectors.
- • GPU cluster (3+ nodes, A100 class)
- • 500+ concurrent seats
- • Air-gapped (fully offline) certified
- • 24×7 SLA + on-site engineer
- • BDDK / KVKK / public-sector audit reporting
Hardware & certification.
Recommended Hardware
- • 16+ core CPU (AMD EPYC or Intel Xeon)
- • 64 GB RAM (128 GB recommended for scale)
- • NVIDIA GPU: RTX 4090 / A6000 / A100
- • 2 TB NVMe SSD
- • Compatible with 1U/2U rack
Compliance & Certification
- ✓ KVKK compliant, data residency in Turkey
- ✓ ISO 27001 aligned processes
- ✓ Ready for BDDK audit reporting
- ✓ Audit trail + WORM-compliant logs
- ✓ Domestic software — national scheme eligible