helix-foundry is a free, open source data engineering & integration project written in TypeScript and released under Apache-2.0. It has 1,680 GitHub stars, 257 forks and 8 open issues, and was last pushed 14 hours ago. On this registry it ranks #28 of 53 tracked projects in Data Engineering & Integration, with 5 head-to-head comparisons available.

What is helix-foundry?

Helix Foundry is an Apache-2.0 data integration and knowledge-graph application that connects a company's data sources, has an AI agent draft an ontology over them, and loads the result into HelixDB as one connected graph that people can explore and ask cited questions of, available hosted at foundry.helix-db.com or on a own machine via Docker Compose.

What it is

Helix Foundry sits in the Data & Analytics / Data Engineering & Integration category. Its front end is a Next.js site, application and same-origin API proxy at the repository root; its backend is the Go module github.com/helixdb/helix-foundry/backend, which contains the API, the workers that sync data, draft ontologies, load them into HelixDB and answer questions, and their MySQL store. Everything in the repository, the site and the backend is Apache-2.0, and so are the three images it builds: foundry-site and foundry from this repository, and ontology-agent, the ontology agent's image.

The concrete problem it addresses is the gap between a company's operational data and a graph anyone can query. Instead of hand-designing a schema and writing the ETL to fill it, an operator connects sources such as PostgreSQL, MySQL, Amazon S3, Stripe, PostHog, WorkOS, GitHub, Linear, Granola and Cal.com, or uploads documents in Markdown, text or PDF alongside table files in CSV, TSV, JSON, JSON Lines or Parquet. An agent then drafts the ontology itself: the types, the links, the tables and documents each comes from, and rules. A person reviews and edits that draft, makes a version current, and a load writes the version into HelixDB as one connected graph.

Key capabilities

  • Connectors for PostgreSQL, MySQL, Amazon S3, Stripe, PostHog, WorkOS, GitHub, Linear, Granola and Cal.com, plus document uploads in Markdown, text and PDF and table files in CSV, TSV, JSON, JSON Lines and Parquet.
  • An agent drafts the ontology — types, links, the source tables and documents behind each one, and rules — for a person to review and edit.
  • Ontology versioning: a reviewed draft is made the current version, and a load writes that version into HelixDB.
  • Exploration of the loaded graph either as a graph or as records, with every value traceable back to its source row or its quote.
  • Question answering over the graph with cited answers.
  • Provider-agnostic AI selection in Settings › AI models, which reports which model was chosen and why, and switches between Ollama, Claude, OpenAI, Azure OpenAI, Amazon Bedrock with your own key, or any OpenAI-compatible server such as vLLM or a gateway.

Who uses it and how

  • Individuals running it locally: docker compose up -d --wait produces one owner, every connector and feature switched on, no sign-in and a workspace to start in at http://localhost:3000.
  • Teams that would rather not operate it themselves use the hosted instance at foundry.helix-db.com, where the hosted Helix runs the same backend code as a library.
  • Organisations consolidating SaaS and database sources such as Stripe, GitHub, Linear, PostHog, WorkOS, Cal.com, PostgreSQL, MySQL and S3 into a single graph for cross-source questions.
  • Analysts and engineers who need provenance, since each value follows back to its row or quote and each answer carries citations.
  • Developers working on the project itself use pnpm dev:mock to run the hosted site against a mock API with no account or secret required, with the layout and rules described in AGENTS.md.

Getting started

Self-hosting requires Docker with Compose v2: clone https://github.com/helixdb/helix-foundry.git, change into helix-foundry, run docker compose up -d --wait and open http://localhost:3000. The hosted option at foundry.helix-db.com needs no installation, and developers can run pnpm install followed by pnpm dev:mock on Node 24 with pnpm.

How it compares

The facts provided name no comparable or paid product that Helix Foundry replaces, so it stands alone in this registry; HelixDB is the graph database it loads into rather than an alternative to it. Its differentiators as documented are the Apache-2.0 licence across the site, backend and all three images, and the choice between the hosted instance and a self-hosted Compose stack.

When to use it — and when not to

A self-hoster must operate the Compose stack, including the MySQL store the backend keeps and the HelixDB instance that local/compose.yaml runs beside the Foundry images, and must supply an AI provider key or a local model such as Ollama or LM Studio. The local install is deliberately single-owner with no sign-in, so it does not suit anyone needing multi-user authentication, and install secrets and data live in a Docker volume that docker compose down -v deletes outright. The top-level README is a quickstart rather than a full operations manual, so operators should read local/README.md, backend/README.md and AGENTS.md before committing.

project readme (upstream, from github) — read inline

Helix Foundry

Connect your company's data, let an agent draft an ontology over it, and load it into HelixDB as one connected graph you can explore and ask questions of.

  • Connect PostgreSQL, MySQL, Amazon S3, Stripe, PostHog, WorkOS, GitHub, Linear, Granola and Cal.com, and upload documents (Markdown, text, PDF) and table files (CSV, TSV, JSON, JSON Lines, Parquet).
  • An agent drafts the ontology: types, links, the tables and documents each comes from, and rules. You review it, edit it and make a version current.
  • A load writes the version into HelixDB. Explore it as a graph or as records, follow every value back to its row or its quote, and ask questions with cited answers.

Use it hosted at foundry.helix-db.com, or run it on your own computer.

Run it on your computer

You need Docker with Compose v2.

git clone https://github.com/helixdb/helix-foundry.git
cd helix-foundry
docker compose up -d --wait

Open http://localhost:3000. There is no sign-in and nothing to set up: one owner, every connector and feature on, and a workspace to start in. The first start writes the install's secrets into a Docker volume; keep it (docker compose down -v deletes it with your data).

AI runs where it finds it: Claude or OpenAI when the shell you ran docker compose up from has an ANTHROPIC_API_KEY or OPENAI_API_KEY, else a model that calls tools in Ollama (docker compose --profile ollama up -d runs one beside Foundry, or use Ollama on your computer), else LM Studio on your computer. Settings › AI models says which it chose and why, and switches to Ollama, Claude, OpenAI, Azure OpenAI or Amazon Bedrock with your own key, or any OpenAI-compatible server (vLLM, a gateway). local/README.md says what runs, where data lives, what may be overridden, how each provider is set up and how to reach databases on this machine.

Develop

The site, the Foundry app and its same-origin API proxy (Next.js), is at the root. It needs Node 24 and pnpm.

pnpm install
pnpm dev:mock        # the hosted site against a mock API, on http://localhost:3000
pnpm dev:mock:local  # the local build against the mock

The mock stands in for the API and sign-in, so no account or secret is needed. Before sending a change, run pnpm typecheck, pnpm lint, pnpm test and pnpm e2e (Playwright; needs pnpm exec playwright install chromium). AGENTS.md describes the layout and the rules.

Backend

backend/ is Foundry's backend, the Go module github.com/helixdb/helix-foundry/backend: the API, the workers that sync data, draft ontologies, load them into HelixDB and answer questions, and their MySQL store. backend/README.md says how to build, test and run it. Hosted Helix runs the same code as a library.

License

Everything in this repository, the site and the backend, is Apache 2.0, and so are Foundry's three images: foundry-site and foundry, built from it, and ontology-agent, the ontology agent's. HelixDB, which local/compose.yaml runs beside them, is Apache 2.0 too.

Frequently asked questions

Is helix-foundry free to use?

helix-foundry is open source under the Apache-2.0 licence. There is no licence fee and no seat count — you can self-host it or, where the project offers one, pay a vendor for a managed version instead.

What does helix-foundry do?

Connect your company's data, let an agent draft an ontology over it, and load it into HelixDB as one graph you can explore and ask questions of. Hosted at found

What is helix-foundry written in?

helix-foundry is primarily written in TypeScript. Its source is publicly available at https://github.com/HelixDB/helix-foundry, and it has 1,680 GitHub stars.