bagofwords1/bagofwords

Chat with your data - with memory, rules, and observability built in. Deploy in 2 minutes

What it solves

Bag of Words (BOW) solves the challenge of connecting LLMs to fragmented enterprise data and tools while maintaining governance. It prevents the need to rebuild agent logic for different interfaces by providing a centralized platform where agents are configured with specific data, credentials, and instructions, then deployed across various channels like chat, dashboards, and automated workflows.

How it works

BOW acts as a governed gateway between LLMs, data sources, and output channels. It allows administrators to define an "analysis context" for each agent—specifying which databases, APIs, or files it can access and what instructions it should follow. It includes a query acceleration layer that caches analytical queries to reduce latency and cost. To ensure reliability, it uses an evaluation system where agents can enter self-improving loops: if a predefined eval fails, the agent drafts instruction fixes and re-tests them until they pass.

Who it’s for

It is designed for enterprise teams and data analysts who need to build and scale reliable AI agents for data analysis, root cause investigation, and automated reporting across diverse data environments (SQL warehouses, BI tools, business apps).

Highlights

  • Multi-Channel Deployment: Run the same agent in a web app, Slack, Microsoft Teams, WhatsApp, Excel, or via MCP clients like Claude Code.
  • Extensive Connectivity: Native connectors for dozens of databases (Snowflake, BigQuery, PostgreSQL), BI tools (Tableau, Power BI), and business apps (Salesforce, Jira).
  • Self-Improving Agents: Automated loops that use evals to detect failures and draft their own instruction updates to fix them.
  • Query Acceleration: An auto-updating cache that can speed up analytical queries by 10–1000×.
  • MCP Gateway: Integrates with Model Context Protocol (MCP) servers and custom APIs to expose tools to MCP clients.

Related

  • Project
    getnao/nao

    An open-source framework to build and deploy analytics agents that enable business users to query data using natural language and generate visualizations.

  • Project
    lightdash/lightdash

    An open-source agentic BI platform that uses a governed context layer to power dashboards, AI agents, and data apps with a software-engineering workflow.

  • Project
    eosphoros-ai/DB-GPT

    An open-source agentic AI data assistant that connects to databases and files to autonomously write SQL and code for data analysis and reporting.

  • Project
    AI4Finance-Foundation/FinRobot

    FinRobot is an open‑source, multi‑agent AI platform for high‑stakes financial analysis. It separates deterministic Python‑based valuation (DCF, LBO, etc.) from LLM‑generated narration, orchestrates specialized agents (data, analysis, modeling, synthesis, reporting, plus debate agents), and ships three versions: a production desktop app (V2), a self‑hosted web report generator (V1), and an educational AutoGen library (V0). The system integrates multiple market data sources, provides 56 analyst playbooks, and can be run locally via a CLI, a browser UI, or a native Tauri desktop client.

  • Project
    microsoft/agent-framework

    Microsoft Agent Framework is an open‑source SDK for building production‑grade AI agents and multi‑agent workflows in Python, .NET, and Go. It offers provider‑agnostic LLM access, graph‑based orchestration, middleware, OpenTelemetry observability, declarative YAML agents, skill integration, and easy hosting on Azure or Microsoft Foundry.