The latest articles on DEV Community by Alex (@alexander_abakah_205cc567). https://dev.to/alexander_abakah_205cc567 https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3976962%2F1a3730c3-8f25-4dba-8be4-bbd5a54a0c1d.png https://dev.to/alexander_abakah_205cc567 en Alex Wed, 10 Jun 2026 05:30:44 +0000 https://dev.to/alexander_abakah_205cc567/intentql-a-semantic-compiler-for-natural-language-postgres-analytics-184f https://dev.to/alexander_abakah_205cc567/intentql-a-semantic-compiler-for-natural-language-postgres-analytics-184f <p>I just released <strong>GroundedQL 0.3.0</strong>, an open-source semantic compiler for reliable natural-language analytics over Postgres.</p> <h2> The problem </h2> <p>Most text-to-SQL systems allow a language model to generate executable SQL directly. When a query fails, it can be difficult to determine whether the problem came from language interpretation, schema selection, joins, SQL generation, or safety.</p> <h2> The approach </h2> <p>GroundedQL separates interpretation from execution:</p> <ol> <li> <strong>Model</strong> → proposes lightweight semantic hints such as datasets, filters, dimensions, and metrics</li> <li> <strong>Compiler</strong> → resolves those hints against an allowlisted <code>schema.yaml</code>, builds and validates a typed <code>QueryPlan</code>, then compiles parameterized SQL</li> </ol> <p>The model interprets language. The compiler owns execution.</p> <p>Filter values become bind parameters rather than string-concatenated SQL.</p> <h2> Quick start </h2> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>pip <span class="nb">install </span>groundedql groundedql init <span class="nt">--db</span> <span class="s2">"postgresql://user:pass@host/db"</span> </code></pre> </div> <div class="highlight js-code-highlight"> <pre class="highlight python"><code><span class="kn">from</span> <span class="n">sqlalchemy</span> <span class="kn">import</span> <span class="n">create_engine</span> <span class="kn">from</span> <span class="n">groundedql.agent</span> <span class="kn">import</span> <span class="n">QueryAgent</span> <span class="n">agent</span> <span class="o">=</span> <span class="nc">QueryAgent</span><span class="p">(</span> <span class="n">engine</span><span class="o">=</span><span class="nf">create_engine</span><span class="p">(</span><span class="sh">"</span><span class="s">postgresql+psycopg2://user:pass@host/db</span><span class="sh">"</span><span class="p">),</span> <span class="n">schema_path</span><span class="o">=</span><span class="sh">"</span><span class="s">config/schema.yaml</span><span class="sh">"</span><span class="p">,</span> <span class="n">llm</span><span class="o">=</span><span class="sh">"</span><span class="s">mistral</span><span class="sh">"</span><span class="p">,</span> <span class="p">)</span> <span class="n">result</span> <span class="o">=</span> <span class="n">agent</span><span class="p">.</span><span class="nf">ask</span><span class="p">(</span><span class="sh">"</span><span class="s">Which customers placed the most orders last year?</span><span class="sh">"</span><span class="p">)</span> <span class="nf">print</span><span class="p">(</span><span class="n">result</span><span class="p">[</span><span class="sh">"</span><span class="s">rows</span><span class="sh">"</span><span class="p">])</span> <span class="nf">print</span><span class="p">(</span><span class="n">result</span><span class="p">[</span><span class="sh">"</span><span class="s">sql</span><span class="sh">"</span><span class="p">])</span> </code></pre> </div> <p>Optional extras:<br> </p> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>pip <span class="nb">install</span> <span class="s2">"groundedql[memory]"</span> pip <span class="nb">install</span> <span class="s2">"groundedql[openai]"</span> </code></pre> </div> <h2> What GroundedQL provides </h2> <ul> <li>Schema allowlisting through <code>schema.yaml</code> </li> <li>A typed intermediate representation called <code>QueryPlan</code> </li> <li>Deterministic validation and SQL compilation</li> <li>Parameterized Postgres SQL through SQLAlchemy</li> <li>Support for Mistral, Ollama, and OpenAI-compatible providers</li> <li>Inspectable plans that can be tested independently of the model</li> </ul> <p>GroundedQL is licensed under <strong>Apache License 2.0</strong>. Contributions, feedback, and real-world testing are welcome.</p> <h2> Links </h2> <ul> <li> <strong>Documentation:</strong> <a href="https://certifore.github.io/groundedql_docs/" rel="noopener noreferrer">https://certifore.github.io/groundedql_docs/</a> </li> <li> <strong>GitHub:</strong> <a href="https://github.com/Certifore/groundedql" rel="noopener noreferrer">https://github.com/Certifore/groundedql</a> </li> <li> <strong>PyPI:</strong> <a href="https://pypi.org/project/groundedql/" rel="noopener noreferrer">https://pypi.org/project/groundedql/</a> </li> <li> <strong>Release:</strong> <a href="https://github.com/Certifore/groundedql/releases/tag/v0.3.0" rel="noopener noreferrer">https://github.com/Certifore/groundedql/releases/tag/v0.3.0</a> </li> <li> <strong>Discussions:</strong> <a href="https://github.com/Certifore/groundedql/discussions" rel="noopener noreferrer">https://github.com/Certifore/groundedql/discussions</a> </li> </ul> <p>If you are building natural-language analytics over Postgres and want the model out of the SQL driver's seat, I would love your feedback.</p> python postgres opensource llm