memfor
Theme
Motion
git clone https://github.com/r5rana/agentware.git
cd agentware
git checkout 3c25cdd
claude "hi"
git clone https://github.com/r5rana/agentware.git
cd agentware
git checkout 3c25cdd
codex "hi"

This installs agentware, the open-source engine behind memfor. No tagged release yet. These commands pin commit 3c25cdd.

Apache-2.0. No memfor account. Needs macOS or Linux (Windows via WSL), git, Python 3, bash, jq, and Claude Code or Codex.

Send setup steps to my computer Join the Teams waitlist

Brand. Names, marks and words, for anyone writing about memfor.

Naming

Naming
memfor The company and master brand. Lowercase, one word.
mem for AI mem for Code mem for Teams Early access · not built mem for Learning Exploring Product lines. Lowercase mem for plus the audience, capitalised.
agentware The open-source engine. Lowercase.
Don't write MemFor, Memfor AI, mem4, Mem on its own, memfor app, the memory layer.

The Index Cell

The mark.

Five lit cells because recall returns the top five. Drawn in rank order, the path makes an M.

Paper
256 px
64 px
24 px
16 px
Graphite
256 px
64 px
24 px
16 px
Glass
256 px
64 px
24 px
16 px

Downloads

Mark, wordmark and lockups as SVG and PNG (light and dark).
Asset LightDark
Mark SVG PNG SVG PNG
Wordmark SVG PNG SVG PNG
Lockup SVG PNG SVG PNG

Colour

  • Ledger paper #F3F5EF
  • Ink #15191F
  • Graphite #121417
  • Instrument glass #07090B
  • Recall yellow #FFE45C
  • Highlighter #FFEE8A

Yellow only marks what was recalled or measured.

Type

Do / don't

Do

  • Use the one-liner.
  • Link numbers to their receipts.
  • Show the Teams chip. Early access · not built

Don't

  • Recolour the mark.
  • Put it on yellow.
  • Add glows.
  • Use #1, best or SOTA.

Boilerplate.

Short
Deterministic, local memory for AI agents.
Long
memfor is open-source memory for AI agents such as Claude Code and Codex. Knowledge lives as plain text in a folder you own, the memory toolkit never calls an LLM, and every benchmark number links to the ledger row that produced it.

Founder bio

Rahul Rana is a security-first full-stack and AI-agent engineer based in Dublin, Ireland. He builds agentware, an open-source memory and work loop for AI coding agents, and memfor, its home.

Press contact: hello@memfor.ai.

Esc