Troubleshooting Your Vector Database: 5 Points to Check Before Tuning Qdrant
Stop wasting time tuning parameters. Learn the 5 critical steps for troubleshooting your vector database to fix retrieval errors and improve Qdrant accu…
Blog
Notes from the stack
Notes, tutorials, and operating patterns from the d0a1 stack.
-
Read more: Troubleshooting Your Vector Database: 5 Points to Check Before Tuning Qdrant
-
Self-Hosted AI in WordPress with Ollama and Qdrant
Read more: Self-Hosted AI in WordPress with Ollama and QdrantLearn how to implement self-hosted AI in WordPress with Ollama and Qdrant to keep your data private, avoid provider quotas, and control your AI infrastr…
-
Measuring RAG Quality: From Retrieval Metrics to Generation Evaluation
Read more: Measuring RAG Quality: From Retrieval Metrics to Generation EvaluationLearn the essentials of measuring RAG quality, from optimizing retrieval and using MTEB leaderboards to implementing cross-encoders for better re-ranking.
-
Putting MCP in Production: Permissions, Quotas, and Logs
Read more: Putting MCP in Production: Permissions, Quotas, and LogsLearn the essential layers for putting MCP in production. Secure your AI agents with robust permissions, usage quotas, and logging for operational matur…
-
Local AI and GDPR: Privacy by Design in Your Own Infrastructure
Read more: Local AI and GDPR: Privacy by Design in Your Own InfrastructureLearn how Local AI and GDPR work together to ensure privacy by design. Discover how hosting models on your own infrastructure simplifies data compliance.
-
Open-Weight Generative Video: Wan, LTX-Video, and Mochi
Read more: Open-Weight Generative Video: Wan, LTX-Video, and MochiExplore the landscape of open-weight generative video with Wan, LTX-Video, and Mochi. Learn about VRAM requirements and how to run them locally via Comf…
-
DeepSeek-R1 has 671 billion parameters and an MIT license
Read more: DeepSeek-R1 has 671 billion parameters and an MIT licenseDeepSeek-R1 has 671 billion parameters and an MIT license, but its massive size makes local hosting nearly impossible due to extreme VRAM requirements.
-
FLUX.1 [dev] and ComfyUI: Local Image Generation for Professionals
Read more: FLUX.1 [dev] and ComfyUI: Local Image Generation for ProfessionalsMaster professional local image generation with FLUX.1 [dev] and ComfyUI. Learn how to use node-based workflows for precise, reproducible AI art creation.
-
Local AI Bot on your VPS: Privacy and Isolation with Telegram and Ollama
Read more: Local AI Bot on your VPS: Privacy and Isolation with Telegram and OllamaA Telegram bot backed by Ollama on your own server: long polling, Docker Compose, per-chat memory and rate limits. No conversation leaves your VPS.
-
Connecting Local AI: Building a Tool-Enabled Stack with MCP, Ollama, Qdrant, and WordPress
Read more: Connecting Local AI: Building a Tool-Enabled Stack with MCP, Ollama, Qdrant, and WordPressBuild an AI agent that taps Ollama models, Qdrant memory, and your WordPress REST API through MCP, with a Docker Compose walkthrough and security tips.
-
Set Up Your Own Private AI Chat with Open WebUI and Ollama
Read more: Set Up Your Own Private AI Chat with Open WebUI and OllamaSet up Open WebUI on Ollama and chat with local models from your browser. Your data stays on your machine, and nobody logs your prompts.
-
Three Common Mistakes When Implementing RAG in WordPress
Read more: Three Common Mistakes When Implementing RAG in WordPressYou launch an assistant based on the 2,000 articles on your site, and the first real question fails: the bot can’t find a guide published two years ago. The culprit is rarely the model; the…