NTT DATA / Innovation Center Brazil

Translate PowerPoint decks with Azure-backed DeckFlow.

Upload a `.ppt` or `.pptx`, choose the source and target language, follow the runtime stage, and download the translated deck from the same workspace.

Azure OpenAI

Configured through environment variables inside the container.

DeckFlow queue

Each translation keeps its own status line and downloadable output.

deckflow-console://job-runner

Translator

Start a translation run

Checking config

Use auto-detect when the source language varies.

Choose a different target before starting the job.

Track current job

Maximum upload size: checking runtime config.

Current job

No active translation yet

Idle

Upload a deck to start the DeckFlow pipeline.

Download translated file

Workflow

The runtime stays explicit from intake to export.

  1. 1.0

    Intake.

    The upload surface stores one deck per job and binds the chosen language pair.

  2. 2.0

    Translate.

    DeckFlow uses the Azure OpenAI deployment already configured in the environment.

  3. 3.0

    Export.

    The completed deck is exposed as a direct download from the same job timeline.

Runtime state

Follow each queue item without leaving the console.

The dark band carries the live queue: current stage, translated output availability, and recent logs emitted by the DeckFlow runtime.

Recent jobs

0 jobs

No translation job has been created yet.

Formats and rules

Operational constraints stay visible.

Accepted inputs .ppt and .pptx validated before the job enters the queue
Maximum upload Checking runtime config taken directly from the server environment
Output Translated PowerPoint deck downloaded directly from the finished job
Language pairing Source + target source can use auto-detect
Execution model FastAPI + background worker one self-contained Docker service
Translator engine DeckFlow runtime using the Azure OpenAI translation pipeline behind the console