Lilac
Archived open-source dataset curation tool, acquired by Databricks
Overview
Lilac was an open-source tool for clustering, searching and cleaning text datasets used to train and evaluate LLMs. Databricks acquired it in March 2024 and archived the repository on 25 July 2025. The Apache-2.0 code remains on GitHub and PyPI, but its website, documentation and hosted Garden service are gone.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 12 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
Use cases
What you can produce with Lilac
- In the archived code: clustering of large text datasets with auto-generated cluster titles
- PII scanning of training or fine-tuning corpora
- Near-duplicate detection and filtering
- Semantic and keyword search through a local browser UI
- Custom concept tagging and export of filtered subsets
- Text statistics and quality signals through the Python API
Lilac made large text datasets explorable: it clustered and auto-titled documents, ran semantic and keyword search, and flagged PII and near-duplicates through a local UI and Python API. Databricks acquired it in March 2024 and archived the repository in July 2025.
Treat it as reference code, not a product. The package is still on PyPI, but the hosted service, docs and demo no longer work, and the old lilacml.com domain and lilacai GitHub address now belong to unrelated parties; use only github.com/databricks/lilac.
What we found
Lilac is finished as a product. Databricks announced on 19 March 2024 that the Boston start-up was joining it, with the technology headed for its Mosaic AI tooling and the terms undisclosed. The open-source repository, now at github.com/databricks/lilac, was archived and made read-only on 25 July 2025. The last release, 0.3.9, is dated 29 February 2024, and nothing has been pushed since March 2024.
Two later changes matter more than the archive flag. The lilacml.com domain no longer belongs to the project: archived copies show a German sports-betting page in May 2026 and a redirect to an unrelated German site in June, and today it returns a Cloudflare block. The old github.com/lilacai address, which the PyPI listing still gives as the homepage, now holds a placeholder repository created by an unrelated account. Neither should be trusted for code or downloads; the Databricks-owned repository is the only reliable source. Earlier versions of this page linked the old domain and said the documentation was still online.
What remains is usable for one-off local work. The Apache-2.0 package is on PyPI, and the code clusters and auto-titles documents, runs semantic and keyword search, and computes signals for PII, near-duplicates and language. The hosted Lilac Garden service, which the team said could cluster and title a million data points in 20 minutes, went with the website, and the public Hugging Face demo now shows a runtime error. We have not test-installed the package, and its 2024 dependency pins may need work on a current Python.
During its year of active development Lilac was well regarded: its README names Cohere and Databricks as users, and the team published clustered versions of datasets such as OpenOrca, UltraChat and LMSYS-Chat-1M on Hugging Face.
Databricks has not announced a named product built from Lilac, so its value today is as reference code for dataset curation.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including Lilac's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Lilac unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Lilac directly →
Spotted something out of date? Suggest an update →
Related reading
Alternatives to Lilac
More in Code & Dev Tools
