Everything (reference) vs Engram
Both are memory & reasoning MCP servers, and we ran both through the same test (recall across sessions) in each app. Everything (reference) passed 2 of 2 apps we tested it in; Engram passed 2 of 4.
Test results side by side
| Everything (reference) | Engram | |
|---|---|---|
| Claude Code | Passed 2026-10-13 | Partial 2026-10-15 |
| Cursor | Passed 2026-10-13 | Partial 2026-10-15 |
| Claude Desktop | Not tested yet | Passed 2026-10-21 |
| ChatGPT (Codex) | Not tested yet | Passed 2026-10-21 |
| Tools exposed (protocol check) | 13 tools | 22 tools |
| GitHub stars | 91,266 (shared repo) | 7,201 |
| License | NOASSERTION | MIT |
| Last push | 2026-10-13 | 2026-10-14 |
Results from our own harness runs, dated per cell. Stars and last push from GitHub at our last data refresh. We don’t pick a winner here: which one fits depends on your app and setup.
What each does
Everything (reference): Reference test server from the MCP maintainers that exercises every protocol feature (tools, prompts, resources). Useful for checking a client, not for daily work.
Engram: Persistent memory for coding agents: a Go binary storing memories in SQLite with full-text search over MCP.
