What Happens Inside an OCR API Call - OCR Fundamentals for Backend Developers
OCR finds text, then reads it. From classic pipelines to deep learning, masking a contract with CLOVA OCR before a VLM, and how a VLM differs from OCR.
llm
5 posts
LLM API Gateway: OpenRouter, LiteLLM, and Cloudflare
An LLM API gateway puts OpenAI, Anthropic, and Gemini calls in one place. Compare OpenRouter, LiteLLM, and Cloudflare on latency, fallback, cache, and budgets.
When AI Caching Pays Off
An AI API cache hit reuses input-prefix KV, not a stored answer. Which requests hit, the break-even rate after write cost, and agent versus chatbot logs.
Monitoring a LiteLLM Proxy - What the Admin UI Covers, and What Only Grafana and Langfuse Show
Prometheus, Grafana, and Langfuse on a LiteLLM proxy, same mock traffic for 14 minutes. What the Admin UI shows, and when to add the other two.
MCP Servers Were Eating 6.7GB of My RAM, So I Built a Proxy
Per-session stdio MCP servers reached 169 processes and 6.7GB. A proxy shares one process, then asks whether these servers still need to be MCP.