2 deep-dives covering this topic
Retrieval-Augmented Generation (RAG) enhances LLM outputs by fetching relevant external documents from vector stores before generating answers.
Persistent memory for AI coding agents
Memfog quietly captures what Claude Code, Cursor, Codex, Gemini CLI, OpenCode and Cline do in every session — prompts, file edits, terminal output, decisions — then surfaces the right context next time. Drop-in hooks. Sub-second BM25 search. Per-user isolation. Apache-licensed.
Web infrastructure for AI agents — 212 endpoints, one key, MCP-native
Ollagraph is the web-intelligence layer for AI agents: 212 endpoints behind one API key and one MCP server, covering smart crawling, structured extraction, agent observability, AEO audits, DNS and SSL intelligence, and async jobs. Credit-metered, refund-on-fail, built for production.
