# Winston — Historical Intelligence Archive

> Winston is an AI-powered research interface over the Focal Point Publications (FPP) archive: 153,000+ embedded document chunks spanning books, diaries, trial transcripts, expert reports, correspondence, signals intelligence, and legal documents related to the work of historian David Irving.

## About

- **URL**: https://ask.winston.study
- **Archive source**: [Focal Point Publications](https://fpp.co.uk)
- **Document count**: 153,216 embedded chunks
- **Built by**: [irving.studio](https://irving.studio)

## How It Works

Winston uses a Retrieval-Augmented Generation (RAG) pipeline:

1. User submits a natural language query
2. The system searches a ChromaDB vector database of 153,216 document chunks
3. Relevant sources are retrieved, ranked, and deduplicated
4. A language model generates an answer grounded in the retrieved sources
5. Sources are cited inline with clickable references to original documents

## Available Models

- **Gemma 3 12B** (Local) — Free, runs on local GPU
- **Claude Sonnet 4** (API) — ~$0.02/query, Anthropic
- **Claude Opus 4** (API) — ~$0.07/query, Anthropic
- **Claude Haiku 3.5** (API) — ~$0.003/query, Anthropic

## Document Types in the Archive

- Books and manuscripts (Churchill's War, The Destruction of Dresden, Hitler's War, etc.)
- Trial transcripts (Irving v Lipstadt, Nuremberg)
- Personal diaries (1978–2007)
- Expert reports (Evans Report, Longerich Report)
- Signals intelligence (CSDIC transcripts, Ultra decrypts)
- Correspondence and letters
- Articles and website posts from fpp.co.uk

## API

- `GET /api/models` — List available models
- `POST /api/query/stream` — Streaming query endpoint (body: `{"question": "...", "model": "local|sonnet|opus|haiku"}`, returns SSE with sources and token events)
- `GET /api/health` — Health check

## Important Notes

- This is an AI research tool. All answers should be verified against primary sources.
- The archive reflects the published works and collected documents of David Irving.
- Sources are retrieved from embedded document chunks, not generated.

## Links

- [FPP Archive](https://fpp.co.uk)
- [irving.studio](https://irving.studio)
- [Winston Terminal](https://ask.winston.study)

---
© Focal Point Publications. All rights reserved. Built by irving.studio.