Agent memory needs its OpenTelemetry moment

Every agent tool stores memory in its own shape, and every week someone builds another one. None of them can read the others. Observability had the same problem until OpenTelemetry gave everyone one data model to share. A tour of today's memory formats, the proposals that already exist, and what a shared spec could look like.

Mads Thines·November 7, 2026·7 min read

Count the memory formats on your machine

List every place your coding tools keep something they "remember". A typical setup today:

  • an AGENTS.md or CLAUDE.md in most repos, and .cursor/rules/*.mdc in a few
  • an agent writing files to a /memories directory through Anthropic's memory tool
  • Copilot keeping facts about each repo on GitHub's side
  • a memory MCP server keeping a JSONL knowledge graph