Skip to content
Cogni-Code

Introduction

Every time you start a new session, your agent starts from zero. It forgets your preferences, your decisions, the bug you spent two hours on yesterday, the deployment strategy you settled on last week. Cogni-Code fixes that.

It is a persistent, inspectable memory system for Claude Code, Codex CLI, OpenCode, pi, and any MCP-compatible agent. Your agent learns how you work across sessions and gets sharper every time you use it. Memory lives as plain files on your disk — you can read it, edit it, diff it, and back it up with git.

What makes it different

Section titled “What makes it different”

Most agent memory falls into one of three camps:

  • Built-in memory (ChatGPT memory, Claude’s saved context) is opaque, vendor-controlled, and locked to one product.
  • Vector-DB memory (mem0, Letta, Zep) stores embeddings in a database. Powerful retrieval, but you cannot read what your agent knows without a UI.
  • Hand-written context files (CLAUDE.md, AGENT.md, .cursorrules) give you full control but require you to write and maintain them by hand.

Cogni-Code is a fourth option: plain files on disk, maintained automatically, capturing behavior rather than just facts.

Cogni-Code Built-in Vector-DB Hand-written
Where it lives Your filesystem Vendor cloud Database / SaaS Your repo
Can you read it? Yes (cat, grep, diff) Limited UI Partial Yes
Self-hosted Yes No Sometimes Yes
Maintained by Background pipeline Vendor Agent or app You, manually
Captures Behavior + facts Mostly facts Mostly facts Whatever you write
Decays when stale Yes No Usually no No
Multi-agent Yes No Sometimes Manual
Generates its own tools Yes (Skillforge) No No No
Git-backed history Yes No Rarely If you commit it
Section titled “Read next”