Ateme automated multilingual subtitle generation using Vertex AI and Gemini
Ateme is a French video technology company that automates large-scale multilingual subtitle generation for streaming and broadcast customers. The solution replaces a previously manual process that required transcription, translation, timecode spotting, and technical integration. Ateme integrated Google Cloud generative AI into its existing workflow so subtitles can be generated and delivered in minutes.
- Organization
- Ateme
- Industry
- Tech & Comms
- Location
- France
- Published
- January 2024
Reported outcomes
Strategic outcomes
Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks
- Customer
- Ateme
- Provider
- GCP
- Maturity
- Unknown
- Linked source
- Google Cloud Customer Story
No explicit deployment-stage evidence found.
Primary read
Use case focus
Showing 3 of 3
- 1Media processing
- 2Content localization
- 3Accessibility automation
- Generating one hour of subtitles in a given language could take up to 15 hours of manual work.
- The process was costly, fragmented, and difficult to scale across languages.
- Customers needed better accessibility and broader linguistic coverage.
- Ateme added a processing step to its software platform that calls Vertex AI and Gemini models after transcoding.
- The models perform audio transcription, timecode spotting, and subtitle generation.
- Output is reformatted into SRT through an internally developed script and integrated into client workflows.
- The application is containerized as microservices on Google Kubernetes Engine and uses Google Cloud Storage for media files.
- A few minutes are now enough to produce professional-quality subtitles.
- The cost to generate one hour of subtitles is less than a dollar.
- The solution improves access to new linguistic markets and supports regulatory accessibility requirements.
- The workflow is fully automated and scalable for broadcasters and streaming platforms.
Architecture
Ateme's Titan File transcoding application runs as containerized microservices on Google Kubernetes Engine. After transcoding, the system calls Vertex AI and Gemini for transcription, timecode spotting, and subtitle generation. Media files are stored in Google Cloud Storage, and subtitle outputs are reformatted into SRT by an internally developed script before being integrated into the existing client workflow.
Sources & evidence1
- Customer explicitly identified
- Primary source available
- Technical implementation details available
AI-generated summary. Verify important details with the linked sources before relying on this case.
Explore related AI use cases
Was this useful?
Community
Comments
No published comments yet.