Skip to content
forked from finic-ai/finic

Universal APIs for unstructured data. Sync documents from SaaS tools to a SQL or vector database, where they can be easily queried by AI applications

License

Notifications You must be signed in to change notification settings

adrianwedd/psychic

 
 

Repository files navigation

Unified APIs for ingesting unstructured data

Slack Docs License Issues Twitter

Psychic is an open source integration platform to extract and transform unstructured data from SaaS applications like Notion, Slack, Zendesk, Confluence, and Google Drive. Instead of building one integration for each data sources, you can build one integration that works for all data sources, and manage each connection from a GUI. Psychic is designed for startups that use LLMs and vector databases.

There are 4 parts of the platform

  • 🪄 Psychic Link: A modal that lets end users to connect their data sources with a point and click interface, using OAuth when available. Available as an NPM package for React projects, or as a magic link.
  • 🪢 Psychic Connectors: Server-side code that makes API calls and handles token management. New integrations are added by implementing a new DataConnector in the server-side code.
  • 💃 Psychic Models: Universal data models that make it possible to transform unstructured data from disparate data sources into a consistent schema before you receive it. Currently Document and Conversation models are supported.
  • 🎩 Psychic Dashboard: A front-end for the Psychic platform that makes it easy to manage connections and connect data sources without having to add Link to an existing app.

Demo

Using Psychic to connect data from a Notion workspace with a LangChain question and answer app.

Psychic.Twitter.Demo.May.21.2023.1225.PM.mp4

Get an API key to test out the cloud version by creating an account on the Psychic dashboard.

If you have any questions on how to get started, come join our Slack community!.

Roadmap

  • LangChain DataLoader
  • 🚧 CRM integrations (Salesforce, HubSpot, etc)
  • 🚧 Webhook support to notify when source documents/conversations have been updated
  • 🚧 Support for write scopes for each connector, in addition to read scopes

Getting Started - 15 min

Check out the quickstart tutorial to get started.

Contributing

See CONTRIBUTING.md

About

Universal APIs for unstructured data. Sync documents from SaaS tools to a SQL or vector database, where they can be easily queried by AI applications

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • TypeScript 87.0%
  • Python 11.4%
  • CSS 0.7%
  • HTML 0.4%
  • JavaScript 0.3%
  • Dockerfile 0.2%