# Memford

> Shared project memory for AI agents: notes, decisions, reusable standards and version history, accessed through project-scoped MCP connections.

Memford stores selected project knowledge. It does not transfer running processes, repositories or entire conversations between clients. Active reads return active notes and assigned standards. AI consolidation is in evaluation and does not currently replace them with an approved main summary.

Optional terminal hooks support selected checkpoints in Codex and Claude Code on macOS and Linux. Recovery starts from the last confirmed remote checkpoint, not from unsaved work. Client support and setup verification matter.

This is a public product index. It contains no customer notes or credentials and grants no access to private projects.

## Product

- [Memford](https://memford.ai/): Product overview and the setup brief.
- [Shared context](https://memford.ai/shared-context/): Share project notes and decisions across authorized AI agents with Memford. Keep the same project, its sources and the context you choose to save.
- [Project standards](https://memford.ai/project-standards/): Write a rulebook once, assign it across projects and include it in agent context. Memford versions shared standards and supports defaults for new projects.
- [Session continuity](https://memford.ai/session-continuity/): Resume from confirmed project knowledge with Memford. Supported terminal hooks save selected checkpoints and load context at the next session start.
- [Version history](https://memford.ai/version-history/): Inspect original notes, revisions and recorded sources in Memford. Choose what agents read while keeping a dated history of changes to project knowledge.

## Setup and documentation

- [Agent setup guide](https://memford.ai/assets/memory.md): Project selection, authorization and save/read verification.
- [Connect agents](https://docs.memford.ai/docs/connect-agents/): Supported connection paths and setup checks.
- [Shared standards](https://docs.memford.ai/docs/standards/): Assignments, new-project defaults, versions and limits.
- [Notes and history](https://docs.memford.ai/docs/memory-and-history/): Active reads, originals, revisions and provenance.
- [Hooks and recovery](https://docs.memford.ai/docs/hooks/): Activation, checkpoint behavior and recovery boundaries.
- [MCP reference](https://docs.memford.ai/docs/mcp-reference/): Project-scoped tools, resources and change hints.
- [Security](https://docs.memford.ai/docs/security/): Access controls, encryption and provider boundaries.