Assiterr

At its core, our project is a comprehensive and intelligent infrastructure service, purpose-built for Web3 companies and ecosystems. We recognize that in the dynamic and complex world of Web3, traditional static documentation and data management methods fall short in meeting the needs of both developers and internal teams.

Our solution is designed to supercharge these static resources into dynamic, easily queriable assets, essentially creating a smart and adaptable knowledge base. It harnesses the power of a fine-tuned Large Language Model (LLM), to turn complex, unstructured data into meaningful, actionable insights.

For developer relations (DevRel) teams, our solution provides a valuable tool for managing large volumes of developer requests, prioritizing them based on urgency and relevance, and offering quick, accurate responses. It liberates these teams from repetitive, low-value tasks, and empowers them to focus more on strategic activities that enhance developer experience and foster community growth.

For external developers, our platform offers an intuitive interface to access up-to-date, relevant documentation, APIs, and other technical resources. It supports them in keeping pace with the rapid updates in the Web3 landscape, promoting efficiency and creativity in their development efforts.

We believe that this transformation of static resources into a dynamic knowledge base can significantly improve decision-making capabilities, foster stronger community engagement, and drive continuous improvement cycles. Ultimately, our goal is to provide Web3 companies with the tools they need to thrive and innovate in the ever-evolving blockchain landscape.

The project was made using template from anything LLM https://github.com/Mintplex-Labs/anything-llm. We uses Pinecone for semantic similarity seach vectorDB, base on cosine similarity . We use Langchain as the infrastructure wrapper component that enable developers to build out LLM chain service for fetching data collection, process data, storing data, fine tune data and deploying your own large language model Express node service to handle requests and caching that smartly cached already trained document and vections. React for front end application We use rainbowkit for wallet connection We use Airstack to query NFT data, and train user specific data. So user can immediately query related on-chain data when they logged in