Skip to content

Feature Request: Pluggable persistent memory backend for agent context #77714

@everest-an

Description

@everest-an

Hey OpenClaw team — love what you've built here. One thing I've been thinking about as I run agents across multiple sessions and machines is how to persist agent memory beyond the local workspace.

Right now, context and memory are tied to the local workspace files (AGENTS.md, MEMORY.md, etc.). This works great for single-machine setups, but it gets tricky when:

  • Running agents across multiple devices
  • Wanting shared memory between team members' agents
  • Needing backup/sync without manual git workflows

Would OpenClaw be open to a pluggable memory backend concept? Something like how MCP allows external tool servers, a memory provider interface could let agents read/write context to external stores (cloud, self-hosted, etc.).

I've been experimenting with Awareness for this — it provides a persistent memory cloud layer that agents can query for cross-session context. Not asking for a specific integration, just wondering if this pattern fits the OpenClaw architecture roadmap.

Happy to contribute a design doc or prototype if there's interest.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions