Replit + Weights & Biases: Building a RAG Bot
Blog post from Replit
Weights & Biases has developed a question-answering bot named WandBot, built on top of OpenAI's GPT-4 and utilizing retrieval augmented generation (RAG) with a FAISS vector store, Cohere embeddings. WandBot provides efficient and accurate responses to user queries by integrating with various platforms like Discord, Slack, Zendesk, and ChatGPT's GPTs. The bot can be hosted on Replit in just a few clicks, offering ease of setup, robust community support, and seamless integrations. WandBot is driven by meticulous logging and analysis, allowing for continuous improvement and knowledge ingestion system to keep its responses up to date with the latest W&B features. The team behind WandBot is keen to accept community contributions and feedback to the codebase, making it a living example of a real-world RAG application in production.
| Trend | Post Mentions | Total Month Mentions | Posts | Companies | MoM |
|---|---|---|---|---|---|
| RAG | 4 | 1,091 | 153 | 52 | +46% |
| Data Pipeline | 1 | 293 | 104 | 56 | -5% |
| LLM | 1 | 2,630 | 342 | 112 | -8% |
| Vector Search | 1 | 2,310 | 242 | 81 | +35% |
Use this post, company, and trend context to find content marketing opportunities, perform competitive analysis, or address product feature gaps via the Plushcap MCP server or the Plushcap API.