Skip to content

Whisper CPU transcription

AI Platform & Data Hardened AMI

Transcribe English and Spanish recordings on CPU with faster-whisper and a bundled multilingual small model. Responses include text, segments and word timestamps. No model-provider account or GPU is required.

Check the Software Sushi seller catalog for the current offer and availability.

Get started View on AWS Marketplace
AMI release
Whisper CPU 0.1.10 - Ubuntu 24.04
Recommended
t3.medium
Storage
24 GiB+ encrypted gp3
Access
Session Manager or SSH
Released
Oct 18, 2026
Delivery
Amazon Machine Image in your own AWS account

CPU English and Spanish transcription with a multilingual small model.

Text, segment and word timestamps from WAV, FLAC or MP3.

Local processing and private HTTPS access without a model API account.

Quick start

Full guide →
  1. Launch and wait for startup
  2. Use the private connection
  3. Transcribe a recording

Starting configuration

  • t3.medium · 4 GiB minimum
  • 24 GiB root storage
  • Private HTTPS · port 8443

These are minimum starting sizes, not throughput benchmarks.

Scope and limits

Package limits. WAV, FLAC and MP3 with a known duration, up to 10 MiB and five minutes per clip. One request at a time; int8 CPU inference. No diarization, translation, live streaming or voice generation. Accuracy and processing time depend on the recording and instance.

Unresolved security findings

This version has not passed the full Software Sushi security gate. Read the exact version-specific known issues and maintenance, including the raw High/Critical matches and coverage limitations, before deployment. A successful application task or Marketplace listing does not certify a clean host.

Billing

The approved launch software rate is USD 0.02 per running instance-hour for the offer’s listed instance type; confirm the actual offer before subscribing. EC2, EBS and data transfer are billed separately. There is no bundled software trial or annual term. Stop or terminate unused instances to control compute and hourly software usage; retained EBS volumes and other AWS resources can continue to incur charges.

Where to go next

  • Getting started: launch, connect privately and complete your first task.
  • Operations: known issues, backups, upgrades, logs and support.

Upstream project

This is an independently packaged open source product. Software Sushi does not claim upstream or AWS endorsement. faster-whisper and the selected model declare MIT. Media libraries and other dependencies have their own licences; retain the installed notices and corresponding-source materials. Installed source, dependency and licence materials are under /opt/whisper-cpu and /usr/share/doc/whisper-cpu.

Upstream documentation. Software Sushi packages this AMI independently.

Last updated October 17, 2026 · Suggest an improvement