Put your existing hardware to work.

Run call analysis on suitable hardware you already control, including overnight where capacity allows. Keep recordings local and pay no platform subscription.

For your team: The local workflow keeps call content in your environment; Call1 requires no standing access to it.
Audience: QA leads, reviewers and operations teams in 100-agent customer-service contact centers.

Why it matters

You may already have suitable hardware available outside business hours. Use that capacity for recorded-call QA and include the extra power, storage and IT time in your cost estimate.

The working AI path uses Apple Silicon. Check supported models and memory requirements against your workload before choosing a machine. A 100-agent focus describes the customer, not the capacity of one computer.

How it works

01
Check your environment
Choose a processing window and suitable hardware. Review the required model files and their terms.
02
Install and process
Your team runs the local services and uses the review workspace in a browser.
03
Manage access and recovery
Set up accounts, retain recordings according to your policy and assign responsibility for backups and updates.

What’s included

Local AI processing

The current qualified path runs supported models on Apple Silicon. Broader hardware and provider support remains incomplete.

Customer-managed accounts

Use passkey sign-in and reviewer, supervisor and admin roles to control access.

Personal-information masking

The pipeline detects and masks personal information before downstream text analysis. Validate its behavior on your recordings.

Your data

Recordings, transcripts, results and local training data stay in your environment. Your team controls access and retention.

Optional hosting is planned at cost plus 15%. It would use a separate data boundary and require explicit authorization; it does not carry the local-processing promise.

Before you roll out

Measure before sizing

Test audio hours, processing time, storage and enabled analysis. Overnight processing works only when the machine can finish within the window.

Include operating costs

The platform software fee is $0. Hardware, setup, energy, storage, IT and human review can still cost money; add them to the calculator.

Availability

No hardware is included. The local workflow is customer installed and operated.
The source is public, but the core open-source license remains pending. Model terms need separate review.
Hosted and confidential-compute offerings are planned, not available services.

Questions for your team

01
Which suitable machines are available, and for how many hours?
02
Who will install, update and recover the local services?
03
What access, retention and security requirements apply to your calls?