Home

Awesome

HeimdaLLM

Pronounced [ˈhaɪm.dɔl.əm] or HEIM-dall-EM

HeimdaLLM is a robust static analysis framework for validating that LLM-generated structured output is safe. It currently supports SQL.

In simple terms, it helps makes sure that AI won't wreck your systems.

Heimdall Build status Docs GitHub Sponsors PyPI License: Commercial License: AGPL v3 Coverage Status

Consider the following natural-language database query:

how much have i spent renting movies, broken down by month?

From this query (and a little bit of context), an LLM can produce the following SQL query:

SELECT
   strftime('%Y-%m', payment.payment_date) AS month,
   SUM(payment.amount) AS total_amount
FROM payment
JOIN rental ON payment.rental_id=rental.rental_id
JOIN customer ON payment.customer_id=customer.customer_id
WHERE customer.customer_id=:customer_id
GROUP BY month
LIMIT 10;

But how can you ensure the LLM-generated query is safe and that it only accesses authorized data?

HeimdaLLM performs static analysis on the generated SQL to ensure that only certain columns, tables, and functions are used. It also automatically edits the query to add a LIMIT and to remove forbidden columns. Lastly, it ensures that there is a column constraint that would restrict the results to only the user's data.

It does all of this locally, without AI, using good ol' fashioned grammars and parsers:

✅ Ensuring SELECT statement...
✅ Resolving column and table aliases...
✅ Allowlisting selectable columns...
   ✅ Removing 2 forbidden columns...
✅ Ensuring correct row LIMIT exists...
   ✅ Lowering row LIMIT to 10...
✅ Checking JOINed tables and conditions...
✅ Checking required WHERE conditions...
✅ Ensuring query is constrained to requester's identity...
✅ Allowlisting SQL functions...
   ✅ strftime
   ✅ SUM

The validated query can then be executed:

monthtotal_amount
2005-054.99
2005-0622.95
2005-07100.78
2005-0887.82

Want to get started quickly? Go here.

🥽 Safety

I am in the process of organizing an independent security audit of HeimdaLLM. Until this audit is complete, I do not recommend using HeimdaLLM against any production system without a careful risk assessment. These audits are self-funded, so if you will get value from the confidence that they bring, consider sponsoring me or inquire about interest in a commercial license.

To understand some of the potential vulnerabilities, take a look at the attack surface to see the risks and the mitigations.

📚 Database support

There is active development for the other top relational SQL databases. To help me prioritize, please vote on which database you would like to see supported:

Static Badge

📜 License

HeimdaLLM is dual-licensed for open-source or for commercial use.

🤝 Open-source license

The open-source license is AGPLv3, which permits free usage, modification, and distribution, and is appropriate for individual or open-source usage. For commercial usage, AGPLv3 has key obligations that your organization may want to avoid:

📈 Commercial license

The commercial license eliminates the above restrictions, providing flexibility and protection for your business operations. This commercial license is recommended for commercial use. Please inquire about a commerical license here:

License Inquiry