code-session-memory vs engram
Persistent memory for AI coding agents compared
code-session-memory and engram help developers and AI coding agents retrieve useful project context across sessions. code-session-memory indexes conversations from supported tools for semantic search, while engram stores curated project memories with full-text search and several access interfaces.
code-session-memory: Search AI Coding Sessions with OpenAI Embeddings
code-session-memory indexes sessions from OpenCode, Claude Code, Cursor, VS Code, Codex, and Gemini CLI into shared, searchable memory using OpenAI embeddings. It stores data locally with SQLite and sqlite-vec by default, with PostgreSQL and pgvector as an optional shared backend.
engram: Give AI Coding Agents Persistent Memory
Engram is a Go-based memory system that helps AI coding agents retain useful project knowledge between sessions. It stores memories locally in SQLite and makes them available through MCP, a CLI, an HTTP API, and a terminal UI.
| code-session-memory | engram | |
|---|---|---|
| Language | TypeScript | Go |
| License | — | MIT |
| Stars | 20 | 7k |
| Forks | 2 | 724 |
| Last analyzed | Oct 30, 2026 | Oct 3, 2026 |
Key differences
- code-session-memory indexes new messages from supported agent conversations; engram focuses on saving and managing durable project memories rather than storing every conversation turn.
- code-session-memory provides semantic search through MCP or a CLI, plus a web dashboard; engram uses SQLite with FTS5 and offers MCP, CLI, HTTP API, and terminal UI access.
- code-session-memory is written in TypeScript; engram is written in Go. Both use the MIT license.
- code-session-memory defaults to local SQLite and offers PostgreSQL with pgvector for sharing across machines; engram defaults to local SQLite and offers Git Sync or optional Engram Cloud for sharing or replication.
- The listed repositories show 20 stars and 2 forks for code-session-memory, compared with 7k stars and 724 forks for engram; both are unarchived.
Choose code-session-memory if you…
- need to search past conversations semantically across supported coding tools.
- want automatic indexing, filters by tool, project, and date, or session compaction into a restart summary.
- plan to share indexed memory across machines using PostgreSQL and pgvector.
Choose engram if you…
- want agents to retrieve curated project knowledge instead of treating every conversation as a transcript.
- need MCP, CLI, HTTP API, and terminal UI access to project memories.
- want local SQLite storage with Git Sync or optional Cloud replication for sharing.
Compare the numbers side by side, or add more projects
This comparison is generated with AI from the OSRepos analyses of both projects. Always check each project's repository and documentation before choosing.