What problem it solves

Coding-agent notes are commonly tied to one machine, one agent, or a single tool, so important context disappears when work changes hands or continues elsewhere. Teams also need a shared, durable way to preserve decisions, failed approaches, session state, and open questions without depending on opaque storage or an LLM for every operation.

Key features

• Shared memory across agents — Multiple coding-agent harnesses feed one shared memory, allowing the next agent to retrieve prior work and explicit handoffs. Each handoff has an owner and can be claimed exactly once.

  Source: README

• Automatic capture and recall — Lifecycle hooks capture prompts, tool calls, and session boundaries, then turn observations into readable markdown pages. Agents receive a bounded startup brief and can search the stored work.

  Source: README

• Markdown as the source of truth — Memory is stored as ordinary markdown in a git-backed wiki that users can inspect, edit, and synchronize. The derived index can be rebuilt from those files.

  Source: README, crates/ai-memory-wiki

• Zero-LLM operation by default — Capture, search, handoffs, and memory aging can run without an API key. Optional LLM consolidation remains separate from the default path.

  Source: README

• Team and machine continuity — A self-hosted server lets memory follow a project across machines and connect team members' agents while preserving personal attribution and project boundaries.

  Source: README

Quick start

Copied verbatim from the project's README.

curl -fsSL https://github.com/akitaonrails/ai-memory/releases/latest/download/install.sh | sh
ai-memory setup

claude plugin marketplace add akitaonrails/ai-memory && claude plugin install ai-memory@ai-memory
codex plugin marketplace add akitaonrails/ai-memory && codex plugin add ai-memory@ai-memory

Who it may suit

Inference: this may suit developers and teams that switch among coding-agent CLIs, resume work across sessions or machines, and want editable project memory that operates without API access.

Topics to discuss

• How should cross-agent handoffs represent ownership, priority, and unresolved questions?

• Which memory-aging policies are most useful for long-running projects and changing codebases?

• How can teams share durable knowledge while keeping personal sessions and sensitive context separated?

Written by ReadmeX AI from the project's README, source and public GitHub data; it may contain mistakes.

Read the original README https://github.com/akitaonrails/ai-memory#readme

ReadmeX opened this community for a trending project; it is not run by the project. Maintainers can claim it from the community page.