Newsletter
Get notified when new AI tools are added
Join the community.
LMQL (Language Model Query Language) is a programming language for working with LLMs that makes prompts more reliable, modular, and easier to control.
Instead of writing one-off prompt text, LMQL lets you define model queries using templates and types, and apply constraints to the output. This helps teams reuse prompt components, reduce unpredictable responses, and make model behavior more consistent.
LMQL includes a runtime that optimizes how queries are executed against an LLM. This is useful when speed, cost, and generation stability matter—such as in production systems or when building multi-step query chains.
LMQL is a good fit for engineers building LLM-powered features who want a more structured approach than manual prompting: