Apppublic

Lioness256/Sunbird_AI_GenAI_application

sourceHugging Faceupdated 4mo agoView on Hugging Face

0likes

App README

🌻 Sunbird AI Pipeline

A Streamlit web app that takes text or an audio recording, transcribes it, summarises it, translates the summary into an East African language, and reads it back — all powered by Sunbird AI.

Pipeline

Input (text or audio)
  └─► [STT]        /tasks/stt              → transcript
        └─► [LLM]  /tasks/sunflower_simple → summary
              └─► [LLM]  /tasks/sunflower_simple → translation
                    └─► [TTS]  /tasks/tts  → audio

Local setup

bash

git clone https://github.com/Bahati-bk/internship-assessment.git
cd internship-assessment
python -m venv venv
# Windows:
venv\Scripts\activate
# Mac/Linux:
source venv/bin/activate

pip install -r requirements.txt

Copy .env.example to .env and fill in your token:

bash

cp .env.example .env
# edit .env and set SUNBIRD_API_TOKEN=<your token>

Run the app:

bash

streamlit run app.py

Environment variables

Variable	Description
`SUNBIRD_API_TOKEN`	Bearer token for the Sunbird AI API. Get one at https://platform.sunbird.ai

Usage

1.Choose Text or Audio file as input
2.Pick a target language (Luganda, Runyankole, Ateso, Lugbara, or Acholi)
3.Click ▶ Run Pipeline
4.Results appear as each step completes — transcript, summary, translation, and a playable audio clip

Known limitations

—Audio files are capped at 5 minutes
—Supported output languages: Luganda, Runyankole, Ateso, Lugbara, Acholi
—TTS quality depends on the Sunbird AI speaker models for each language