GitHub - omo-ai/omo: Open-source, AI-native enterprise search. Use popular LLMs to chat with your data.

2 min read Original article ↗

Omo chat window

Omo connectors

Omo is an open-source, AI-native enterprise search platform that allows you connect and chat with data sources such as Google Drive, Notion, Confluece, and more. Our goal is to build an LLM and vector store agnostic generative AI search platform that can power search and Q&A use cases such as

  • Internal workplace search
  • In-app search or Q&A
  • External, customer-facing Q&A

Omo is licensed under Apache 2.0.

Documentation & API Reference

Design goals

  1. API-first: support many different clients
  2. Declarative pipelines: flexibility to define and configure indexing and retrieval pipelines steps.
  3. LLM agnostic: plug-in an LLM of your choice
  4. Vector store agnostic: plug-in a vector store of your choice
  5. Cloud-agnostic: Run a self-hosted instance in your cloud
  6. Ensure data is secure and provie ability to host on-prem.

Features / Roadmap

Feature Description Supported Roadmap Notes
Chat UI Turn-based chat UI
Chat history Track recent chat history History length is configurable
Slack client Search via Slack app
Dark mode Toggle between dark and light themes
Source citations Answer returns a list of documents searched
LLM agnostic Support GPT, LLama 3, Claude, or Gemini. Configured via env variable
Embedding model agnostic Use an embedding model of your choice Configured via env variable
Kubernetes / containers Run on container infrastructure AWS EKS supported
Multi-tenancy Single deployment supports many users
Vector store agnostic Use a vector store of your choice Only Pinecone is supported at the moment
Declarative pipelines Define your pipeline steps in YAML
Agents, tools, function calling Low-code UI to create agents
Hybrid search Return ranked list of links in addition to gen AI answer
Search contexts Search personal and org wide Connectors
Multi-modal support Voice and image support

Connectors

The data sources that you can connect are known as Connectors. Supported Connectors include

Connector Supported Roadmap Notes
Google Drive
Notion
Confluence
Airtable
OneDrive
Slack

All connectors are self-service configured via a UI. We'd like to support the widest catalog of Connectors. If you'd like to contribute a connector, please read our contribution guidelines, or if you need a custom connector please contact us.

Contact

If you have general questions, please create a Discussion thread and I'll try to respond ASAP.