Skip to content

Claude Code

Claude Code is the primary integration for Engram. Every time you work with Claude Code, Engram gives it persistent memory across sessions — so it remembers your preferences, your architecture, your debugging insights, and your team’s accumulated knowledge.

Without Engram, every Claude Code session starts from scratch. Your AI has to re-discover your coding conventions, re-read your architecture decisions, and re-learn your preferences. With Engram:

  • Conventions stick — “We use tabs, not spaces” is remembered once and recalled forever
  • Architecture is preserved — Design decisions, API patterns, and system boundaries persist
  • Debugging insights accumulate — Past bug fixes inform future ones
  • Team knowledge flows — Your team’s collective knowledge is available to every engineer’s AI

See the Claude Code Setup guide for installation and configuration.

  1. Claude Code connects to Engram via MCP (Model Context Protocol)
  2. The advise hook injects relevant memories before each prompt
  3. As Claude works, it proactively remembers useful knowledge
  4. Memories are connected into a knowledge graph with typed edges
  5. On team plans, knowledge syncs to the shared team graph

Teams struggle to disseminate conventions, patterns, and decisions to Claude. Each engineer’s Claude session is isolated — there’s no way to share “we always use guard clauses” or “the auth system works like this” across the team.

Engram solves this:

  1. Each engineer’s Claude builds a personal knowledge graph
  2. Project exposure syncs relevant knowledge to the team graph
  3. An AI consolidates team knowledge — deduplicating, resolving conflicts
  4. Every engineer’s Claude recalls from both personal and team graphs

The result: new engineers get the team’s accumulated knowledge from day one, conventions are enforced consistently, and architecture decisions don’t have to be re-explained.

Claude Code has access to all 22 Engram MCP tools. See the Tools Reference for the complete list.