The latest articles on DEV Community by Solo Developer (@solo_developer_ai). https://dev.to/solo_developer_ai 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%2F4236510%2Ff4539243-20e3-46c6-b828-190f399cd31d.jpeg https://dev.to/solo_developer_ai en Solo Developer Tue, 10 Nov 2026 07:37:54 +0000 https://dev.to/solo_developer_ai/i-built-bhasha-kit-a-free-offline-python-library-for-hindi-bengali-text-12h0 https://dev.to/solo_developer_ai/i-built-bhasha-kit-a-free-offline-python-library-for-hindi-bengali-text-12h0 <p>Hi everyone! I'm Solo Developer. I just published my first Python package, <strong>bhasha-kit</strong>, a free, offline toolkit for Hindi, Bengali and English text.</p> <h2> Why I built it </h2> <p>Most text tools need an API key or an internet connection, and many don't handle Hindi and Bengali well. I wanted something simple that runs locally, with no dependencies.</p> <h2> What it does </h2> <ul> <li>Cleans text (emojis, extra spaces, Unicode)</li> <li>Detects Hindi / Bengali / English</li> <li>Transliterates Roman ↔ Devanagari / Bengali</li> <li>Converts numbers to words, including rupees and paise</li> <li>Basic sentiment analysis and keyword extraction</li> </ul> <h2> Install </h2> <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>pip install bhasha-kit </code></pre> </div> <h2> Quick examples </h2> <div class="highlight js-code-highlight"> <pre class="highlight python"><code><span class="k">from</span> <span class="n">bhasha_kit</span> <span class="k">import</span> <span class="n">clean_text</span><span class="p">,</span> <span class="n">number_to_words</span><span class="p">,</span> <span class="n">analyze_sentiment</span> <span class="nf">clean_text</span><span class="p">(</span><span class="s2">"नमस्ते 😀 दोस्तों"</span><span class="p">)</span> <span class="c1"># 'नमस्ते दोस्तों'</span> <span class="nf">number_to_words</span><span class="p">(</span><span class="mi">1250</span><span class="p">,</span> <span class="s2">"hi"</span><span class="p">)</span> <span class="c1"># 'एक हज़ार दो सौ पचास'</span> <span class="nf">number_to_words</span><span class="p">(</span><span class="mi">1250</span><span class="p">,</span> <span class="s2">"bn"</span><span class="p">)</span> <span class="c1"># 'এক হাজার দুই শত পঞ্চাশ'</span> <span class="nf">analyze_sentiment</span><span class="p">(</span><span class="s2">"this is not good"</span><span class="p">)</span> <span class="c1"># {'label': 'negative', 'score': -1.0, 'lang': 'en'}</span> </code></pre> </div> <h2> Known limitations </h2> <ul> <li>Transliteration is rule-based and approximate.</li> <li>Sentiment uses a small word list and doesn't understand sarcasm or context.</li> </ul> <h2> Links </h2> <ul> <li>GitHub: <a href="https://github.com/solodevelopersai/bhasha-kit" rel="noopener noreferrer">https://github.com/solodevelopersai/bhasha-kit</a> </li> <li>PyPI: <a href="https://pypi.org/project/bhasha-kit/" rel="noopener noreferrer">https://pypi.org/project/bhasha-kit/</a> </li> </ul> <p>It's an early version (0.2.0). I'd love your feedback: try it, test it, and tell me about any bugs. A ⭐ on GitHub would help a lot. Thank you!</p> python opensource nlp showdev