Purpose and Demographic
The Translation Editor is a specialized digital workspace designed to support the accurate, efficient, and collaborative translation of Buddhist texts. It provides features like line-by-line translation, AI-generated translations , suggestions, structured exports, and sharing capabilities. The tool is built to respect the traditional formatting of source texts (e.g., Pecha layout) while enhancing the modern translation workflow through digital aids and collaboration tools.
✦ Mission Statement
To empower translators and scholars with a dedicated platform for translating Buddhist texts, combining precision, collaboration, and AI-powered assistance to preserve and share ancient wisdom across languages and cultures.
✦ Target Demographic
- Translators working on classical Buddhist texts across Tibetan, Sanskrit, Pali, and modern languages.
- Translation teams and NGOs collaborating on multi-language publishing and localization of Buddhist literature.
- Monastic institutions involved in teaching, preserving, and translating canonical scriptures.
- Language technologists building digital tools for Tibetan and Buddhist textual ecosystems.
✦ Problem Statement
Translating Buddhist texts presents unique challenges: complex source structures, lack of standardized digital tools, and limited support for collaboration and version control. Traditional workflows are often fragmented, making it difficult to maintain alignment between source and translation, especially in line-sensitive formats like Pechas. The Translation Editor addresses these issues by offering an intelligent, line-aware, and user-friendly platform designed specifically for Buddhist translation work.
Product Objectives
✦ Core Objectives
- Enable seamless para-by-para translation with synced source and target views
- Provide AI-assisted draft generation and suggestion system
- Offer export and share functionalities for easy dissemination
- Support collaboration through comment and annotation features
- Maintain accurate referencing via line number tracking
- Preserve document versions and change history
✦ Non-Goals
- Not intended to replace professional translators with AI
- No OCR or text digitization (handled by other OpenPecha tools)
- Not a full-fledged publishing platform
- This tool can draft a translation , but correction of paragraph sync and translation will not be automated.
✦ Impact Areas
- Make translation easier for translators and scholars
- Empowers global translation efforts with modern tooling
- Supports OpenPecha’s broader mission of enhancing access to Tibetan/Buddhist literature
- Enrich OpenPecha Library of Buddhist text and there translations
Example Use Cases
✦ Use Case: Tsering, a translator
- Loads a Tibetan Pecha and begins translating it line by line
- Uses AI suggestions to accelerate first draft creation
- Discusses specific passages with colleagues via inline comments
✦ Use Case: Emily, a Buddhist studies graduate student
- Compares English and Tibetan versions of a sutra side-by-side
- Uses the suggestion system to refine her translation choices
- Exports the draft as a
.docxfile for her academic advisor
Architectural Considerations
✦ Tech Stack
- Frontend: React, Tailwind, QuillJS
- Backend: FastAPI, PostgreSQL, Express , Prisma, Celery, Redis
- AI: OpenAI/LLM/CLAUDE integration for suggestions and Translations
- Storage: AWS S3 for document storage
- Authentication: Auth0 + role-based access
✦ System Diagram
Workflow
- User submits Document on the React frontend.
- Frontend calls Express backend, which:
- Stores job metadata in Postgres.
- Triggers the translation job by calling FastAPI on EC2.
- FastAPI checks Redis for job result:
- If result exists, returns immediately.
- If not, it queues the job to Celery.
- Celery Worker:
- Calls AI API (Claude/OpenAI) for translation.
- Stores result/status in Redis.
- Frontend polls/checks job status via Express.
- Once complete, Express:
- Fetches the result from Redis.
- Notifies the frontend via WebSocket.
- Translated result is displayed to the user.
✦ Security & Privacy
- User detail and documents are stored on postgres with backend.
- No storage of AI-generated content without user approval.
- User auth will be supported via auth0.
- Token limit for AI translation will be decided its 10000 char for now.
✦ Dependencies
- Frontend + Backend hosted on render.com
- AWS ec2 for translation worker (Docker) [ container with Redis, fastapi, worker]
- OpenPecha libraries and APIs
- QuillJS for rich text editor
- Postgress Database on render.com
✦ Scalability & Maintenance
- Modular design for component update without changing whole application code
- Auto backups and change history/version control system supported by render.com
- Auto Scaling supported by render.com
- EC2 translation worker uses redis for faster response
Participants
✦ Working Group Members
✦ Stakeholders
@Jennifer_Yo [Indrajala]
✦ Point of Contact
- Tenzin Kunsang – [tenkus@esukhia.org]
Project Status
Track the progress of the product over time.
✦ Current Phase
Alpha development – Core features (translation pane, export, AI suggestions) in active build
✦ Milestones
Upcoming goals, releases, or decisions.
- Line-by-line translation + line number sync
- AI translation + suggestion system integration
- Document sharing + export system
- line-by-line docx sharing
- side-by-side docx sharing
- docs with Pecha template containing source and translations
- Collaborative commenting + roles
- Sharing Doc in view-mode with URL
- Realtime Translation generation when working
- Apply footnote similar functionality
✦ Roadmap
| Phase | Features |
|---|---|
| V2 (Q3 2025) | Basic Exports , use Docs Template , Line View, Column View |
| V3 (Q4 2025) | AI Suggestions, Inline Commenting with AI, Versioning with AI , Footnote |
| V4 (Q1 2026) | view-mode Sharing, Real-time Translation with AI |
Meeting Times
✦ Regular Schedule
Every Even Friday at 2:30PM IST via Google Meet
✦ Meeting Notes
What We’re Working On
We maintain a public task board with all active issues and discussions.
