Knowa

• 

  Knowa

Inspiration When new employees join large organizations, they are often overwhelmed by endless forms, policies, and unclear workflows. Managers end up answering the same questions repeatedly, slowing down productivity. We wanted to create a solution that simplifies onboarding and everyday knowledge retrieval, so employees spend less time searching and more time working. What it does Knowa is an AI-powered knowledge assistant that acts as a company’s living memory. It ingests information from websites, documents, apps, and databases. Employees can upload any type of media (text, PDF, image, audio, video) to the Knowa platform, and Supermemory automatically converts it into vector embeddings that feed into the database. It uses RAG (Retrieval-Augmented Generation) to fetch relevant context before generating answers. It leverages Supermemory to remember past interactions, adapt to organizational changes, and continuously update the knowledge base. Employees can interact with Knowa through a web app or Slack integration to get instant, contextual answers about onboarding, workflows, or company policies. How we built it Designed a web interface with login and file upload capabilities for different data formats. Used Supermemory as the ingestion engine to vectorize and store every uploaded file — no matter the format — into the company’s knowledge base. Integrated RAG pipelines so that every query retrieves the most relevant company-specific data before responding. Built persistent memory layers with Supermemory to ensure the assistant grows smarter over time. Connected the system with Slack to provide real-time, conversational support to employees. Challenges we ran into Building end-to-end pipelines that can seamlessly convert any type of data (text, PDFs, images, videos) into a knowledge base connected to Slack. Designing an interface simple enough that even non-technical users can set up and deploy their own Slack bots without needing coding knowledge. Accomplishments that we're proud of Successfully built a working prototype that integrates RAG + Supermemory for both knowledge ingestion and retrieval. Enabled seamless multi-format uploads (documents, media, websites) that automatically convert into vectors. Created a clean, intuitive interface for both web and Slack. Reduced repetitive onboarding questions in our demo setup, showing the real impact of Knowa. What we learned The importance of retrieval quality when working with enterprise-scale data. How RAG dramatically improves accuracy over standard LLM outputs. How Supermemory can serve a dual role: powering knowledge ingestion and enabling contextual recall. The value of designing scalable solutions that adapt as company data grows. What's next for Knowa Expand beyond Slack and web into email and Microsoft Teams integration. Build role-specific knowledge layers so responses adapt to whether the user is a new hire, manager, or HR. Add analytics dashboards to track which questions are asked most often. Scale the system to handle enterprise-level datasets and multilingual support.

Built With

• javascript
• node.js-ai-&-apis:-openai-api
• python
• slack-api-database:-supabase-cloud-&-hosting:-vercel-authentication:-oauth-developer-tools:-github
• supermemory
• typescript-frameworks-&-platforms:-next.js
• vs