Repolode
Submit
HelixDB

HelixDB/helix-foundry

Local-first data workspace that connects company databases, SaaS tools and files into one ontology you can query, running on your own computer.

View on GitHub
Stars
1,617
Forks
243
Open issues
17
Last commit
7 days ago

What is it?

Helix Foundry connects the data sources a company already uses, such as PostgreSQL, MySQL, Neo4j, Stripe, WorkOS, PostHog, REST APIs, S3-compatible storage and uploaded CSV, JSON, JSONL or Parquet files. It proposes a single ontology that matches the same customers and records across these sources, stored as objects and relationships in HelixDB, which you can explore and ask questions of.

The app runs entirely on your machine through Docker Compose, with no accounts or hosted service. Imported data is saved as immutable Parquet snapshots and queried by isolated DuckDB processes. Questions asked in the Analyst view are answered with executed SQL and snapshot citations.

AI runs locally by default using the qwen3:4b model through Ollama, and you can switch to Claude, OpenAI, Azure OpenAI, Amazon Bedrock or a custom OpenAI-compatible endpoint. AI-proposed changes are tested and wait for your review before being applied.

Key features

  • 01Connects PostgreSQL, MySQL, Neo4j, Stripe, WorkOS, PostHog, REST APIs and S3-compatible storage
  • 02Imports CSV, JSON, JSONL and Parquet files, plus an ingestion API
  • 03Suggests a unified ontology that links the same records across sources
  • 04Stores versioned Parquet snapshots with schema-drift review and history
  • 05Answers questions with executed SQL and snapshot citations
  • 06Runs local AI by default through Ollama, with Claude or OpenAI as alternatives

Use cases

  • Getting a combined view of customers across a production database, Stripe and product analytics without a hosted service
  • Exploring how accounts, subscriptions, invoices and support tickets relate to each other
  • Asking natural-language questions about company data while keeping it on a local machine
  • Running a self-hosted data workspace for a team member working alone on sensitive records

Quick start

  1. 1

    Make sure Docker with Compose v2 is installed and running

  2. 2

    Clone the repository

    git clone https://github.com/helixdb/helix-foundry.git
  3. 3

    Enter the project folder

    cd helix-foundry
  4. 4

    Run setup and start the stack with Docker Compose

    ./scripts/setup.sh --compose

Things to know

  • Developed and validated on macOS with Apple Silicon; Windows is untested and the scripts need bash.
  • Requires Docker, and the first run downloads several GB.
  • Local AI benchmarked on a 16 GB machine; a paid Claude or OpenAI key is needed if you prefer cloud AI.

Topics

#data-integration#ontology#duckdb#local-first#knowledge-graph#text-to-sql#self-hosted#docker

FAQ

What is helix-foundry?
Local-first data workspace that connects company databases, SaaS tools and files into one ontology you can query, running on your own computer. Helix Foundry connects the data sources a company already uses, such as PostgreSQL, MySQL, Neo4j, Stripe, WorkOS, PostHog, REST APIs, S3-compatible storage and uploaded CSV, JSON, JSONL or Parquet files. It proposes a single ontology that matches the same customers and records across these sources, stored as objects and relationships in HelixDB, which you can explore and ask questions of.
Is helix-foundry free to use? What license does it have?
helix-foundry is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is helix-foundry still maintained?
Actively maintained. The last commit to helix-foundry was on Oct 13, 2026.
What can I use helix-foundry for?
Getting a combined view of customers across a production database, Stripe and product analytics without a hosted service Exploring how accounts, subscriptions, invoices and support tickets relate to each other Asking natural-language questions about company data while keeping it on a local machine Running a self-hosted data workspace for a team member working alone on sensitive records
How do I install helix-foundry?
Make sure Docker with Compose v2 is installed and running Clone the repository: git clone https://github.com/helixdb/helix-foundry.git Enter the project folder: cd helix-foundry Run setup and start the stack with Docker Compose: ./scripts/setup.sh --compose
Where can I get helix-foundry?
The source code and installation instructions are on GitHub: https://github.com/HelixDB/helix-foundry

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects