LiiiGHTWAVE

LiiiGHTWAVE is reimagining how history is researched, written, and experienced. This AI-powered historiography platform simplifies historical research by allowing users to query and explore curated knowledge bases trained on African and global historical data. By using generative AI, LiiiGHTWAVE makes historical insights accessible to academics, students, and curious minds alike—bridging the gap between dusty archives and digital accessibility.

Historical knowledge is fragmented and hard to access.

Academic research is time-intensive and often siloed

Historical data lacks contextual AI analysis

African history is underrepresented in digital research tools

Oral traditions and non-textual histories are often excluded

LiiiGHTWAVE solves this by building AI tools trained for historical exploration and discovery.

LiiiGHTWAVE currently integrates GPT-4o-mini to generate intelligent responses and summaries based on user prompts. In the long term, it aims to build custom Large Language Models (LLMs) designed for:

Contextual Historical Reasoning

Specialized understanding of historical periods, themes, and figures

Built on a curated corpus of academic and indigenous African knowledge

Multimodal Data Integration

Plans to include oral histories, audio, and visual primary sources

Model will be able to parse and contextualize diverse formats

Ethical AI Training for Cultural Memory

Building frameworks for training on culturally sensitive data

Avoiding historical bias while preserving narrative integrity

Accelerates historical research by automating foundational tasks

Increases access to African history and decolonized perspectives

Supports educators, students, and independent researchers

Democratizes history through AI-driven accessibility

Freemium research tool with subscription for academic/pro features

Partner integrations with schools, universities, and cultural institutions

Licensing model for academic libraries and publishers

Working prototype live using GPT-4o-mini via OpenAI API

Vision to train custom LLMs on curated historical datasets

Early feedback from educators and history enthusiasts

Seeking guidance on copyright-safe training and oral history data integration

Targeting NVIDIA A100 with 40–80GB VRAM

GPU compute needed for future custom model training

Anticipating training cycles for larger, more contextual models

Support our mission by voting in the Hottest African AI Startups competition! Your R20 vote will help preserve and share history using responsible, African-led AI.