Opinion: AI agent project memory doesn't always need vector DB; Markdown+Git is simpler
phucphungbk · reddit · 2026-08-20
A Reddit user argues that for small to medium projects, AI agent memory doesn't need vector DB + RAG; a simpler Markdown+Git approach works. Reasons: vector DBs are hard to inspect, add management overhead, and not all projects need semantic retrieval. He suggests storing memory in .ai-memory/ as Markdown, auditable via Git, no extra infra. He explores the boundary between the two approaches.
More from coding & agent
- Wisp Team Demonstrates Privacy-First AI Workflow with Local Processing and TEE — bgmshana · 2026-08-21
- Refuse cross-session data sharing in Claude settings — dotey · 2026-08-21
- Leaked Stripe Letter Reveals OpenRouter Acquisition, Calls Agents 'Economic Actors' — rohanpaul_ai · 2026-08-21
- Chroma Launches Foundation: Self-Improving Memory for Agents — nbaschez · 2026-08-21
- Epho: Run Claude Code and other coding agents in the cloud via API — karakanb · 2026-08-21
- Designing stopping rules for agents that can always ask for more evidence — ExtremeProgress2201 · 2026-08-21