chroma-core/chroma
Chroma is an open-source search and retrieval database specifically designed for AI applications, supporting document retrieval and vector embeddings.
Awesome AI Agents › RAG and Business Analytics
bagofwords1/bagofwords
Bag of Words is a self-hosted analytics platform in which language model agents answer questions over a company data stack. Each agent is configured individually with the data sources it may read, the tools it may call, its credentials, written instructions and permissions, so different teams can run agents with different scopes on the same deployment. Work starts in a chat interface where an agent plans its steps, writes queries, produces charts, reports and dashboards, and runs deeper investigations such as root cause analysis, reflecting on intermediate results as it goes. The same configured agents can then be reused outside chat: scheduled reports and recurring tasks, event and webhook triggered investigations, and delivery into Microsoft Teams, Slack, WhatsApp, email, Excel and the web application. An MCP gateway works in both directions, letting agents reach external MCP servers and custom APIs, and exposing the platform context and tools to MCP clients such as Claude Code and Codex, so the platform can also be driven headlessly. A caching layer keeps analytical query results fresh and, according to the project, speeds repeated queries by one to three orders of magnitude while reducing load on the source database. Reliability is handled through evaluation sets that describe expected agent behaviour: when an eval fails, a self-improvement loop can draft changes to an agent instruction set and re-run the evals up to a configured limit, with passing candidates either promoted automatically or held for human approval. Governance features include role based access control, approval flows, audit logs, service accounts, single sign-on, model policies and support for row level security, and business definitions can be versioned and synchronised with Git for dbt models, markdown and code. It ships as a Docker image and runs with SQLite or PostgreSQL, with Compose and Kubernetes deployments documented.
https://github.com/bagofwords1/bagofwords
Also in RAG and Business Analytics
vanna-ai/vanna
Vanna is an open-source Python framework that uses Retrieval-Augmented Generation and large language models to convert natural language questions into accurate SQL queries for various databases, simplifying data access and querying.
Canner/WrenAI
Wren AI is an open-source Generative Business Intelligence AI agent that enables data-driven teams to interact with their data using natural language to generate SQL queries, visualizations, and reports for smarter business insights.
import-ai/omnibox
Cross-platform self-hostable AI knowledge hub that collects web pages, files and voice notes, indexes them end to end, and answers questions or drafts text over the local and internet corpus.
apecloud/ApeRAG
Production-ready RAG platform combining graph, vector, and full-text retrieval with AI agents, multimodal document parsing, MCP support, and enterprise deployment features.
newrelic/newrelic-ruby-agent
The New Relic Ruby agent is a comprehensive performance monitoring tool for Ruby applications that helps identify issues, analyze business data, and improve customer experience.
zhongyu09/openchatbi
Open source chat-based business intelligence agent built on LangGraph and LangChain that converts natural language to SQL, runs Python analysis and returns plots, forecasts and anomaly explanations.
superagent-ai/super-rag
Super-Rag is a high-performance RAG pipeline offering summarization, retrieval, reranking, and code interpretation for AI applications through a simple and production-ready REST API.