👉 Try it now for free for 30 days At Regolo, we’re passionate about making cutting-edge AI accessible, compliant, and sustainable. That’s why we’re excited to bring you this updated guide inspired by the latest advancements in Vision-Language Models (VLMs) for Optical Character Recognition (OCR). While open-source VLMs like DeepSeek-OCR are revolutionizing document processing, running […]
Category Archives: Tutorial
Build an Internal Knowledge Bot with Clawdbot: Embeddings + Rerank + Chat in 30 Minutes
👉 Try it now for free for 30 days Build an internal knowledge assistant using Clawdbot that combines semantic search, reranking, and LLM generation on Regolo’s EU infrastructure—with predictable costs and enterprise-grade data privacy. The Knowledge Retrieval Accuracy Problem Most companies accumulate critical tribal knowledge in scattered docs, runbooks, ADRs, and Slack threads—but when developers […]
Orchestrating Predictable AI Agents with Parlant and Regolo
Purpose vs. Predictability Have you ever tried to give an LLM a specific purpose, only to receive unpredictable, poorly structured, or “hallucinated” outputs? Prompt engineering can only take you so far. When building production-ready agents, you need more than just a good prompt; you need a behavioral framework. This is the exact problem Parlant solves. […]
Production-Ready RAG on Open Models: Chunking, Retrieval, Reranking & Evaluation
👉 Build Production RAG on Regolo Time commitment: 20 minutes for setup, 5 minutes per launch after deployment Naive RAG setups chunk blindly, embed with weak models, retrieve irrelevant chunks, and pipe garbage into capable LLMs—resulting in 40% hallucination rates, poor recall, and users abandoning your app after one wrong answer. Build a production RAG […]
Privacy-First Email Search: Building a RAG System with LlamaIndex
👉 Try LLamaIndex on Regolo now Searching through thousands of emails is tedious, slow, and often inaccurate with traditional keyword-based tools. Daniele Scasciafratte‘s demo at the “Build your AI” event in Rome showed a better approach: a privacy-preserving email RAG (Retrieval Augmented Generation) system that indexes emails locally and answers natural language queries without sending […]
From Zero to an Enterprise-Ready AI Agent with Cheshire Cat and Regolo: A Practical Guide Using Only Open‑Source LLMs
Learn how to spin up Cheshire Cat and plug it into Regolo’s OpenAI‑compatible endpoint using open‑source LLMs like Llama‑3.3‑70B‑Instruct, ready for enterprise. 👉 Try CheshireCat and Regolo with Deepseek-r1-70B for free What is CheshireCat AI? CheshireCat is an open-source, production-ready Python framework that allows you to build conversational AI agents with memory, custom tools, and […]
Build Multi-Agent Workflows with crewAI
👉 Try crewAI with Deepseek-r1-70B on Regolo for free What is crewAI? CrewAI is a standalone, lean, and fast Python framework built specifically for orchestrating autonomous AI agents. Unlike frameworks like LangChain, CrewAI does not rely on external dependencies, making it leaner, faster and simpler. Developers end up building fragile chains of sequential API calls, […]
Self-Host With Ease: Using Nextcloud to Create a Smarter Home Setup with regolo
Understanding Nextcloud: The Foundation of Digital Sovereignty Defining Nextcloud Hub: Architecture and Components Nextcloud Hub is fundamentally defined as a self-hosted, open-source content collaboration platform. Its core mandate is to provide users and organizations with a secure environment where they retain full control over their sensitive data, including documents, calendars, contacts, and photos, by storing […]
Qwen3-Reranker-4B on Regolo: Add a “critical brain” to your RAG in minutes
With Qwen3-Reranker-4B, you can turn your retrieval-augmented generation (RAG) system into a smarter assistant: it not only finds relevant documents but also reorders them based on their usefulness for your query. The model uses a cross-encoder that evaluates queries and passages together, ensuring more accurate answers with less noise.
Thunderbird + ThunderAI + Regolo: AI email summaries & replies with an OpenAI‑compatible endpoint
Discover how to enhance Thunderbird with AI using the ThunderAI plugin and regolo.ai/. Learn to configure your API keys, select models like Llama-3.1-8B-Instruct, and leverage AI-powered email features such as summaries, automatic replies, and custom prompts – making your email workflow faster and smarter.