Issue Station
Sign In
Tools hereToolbox

openclaw - 💡(How to fix) Fix Feature: OpenBrain (OB1) MCP as a first-class memory target [1 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
openclaw/openclaw#70262•Fetched 2026-04-23 07:27:03
View on GitHub
Comments
0
Participants
1
Timeline
0
Reactions
0
Author
simonlpaigesimonlpaige
Participants
simonlpaigesimonlpaige

Root Cause

OB1 already ships Claude Desktop, ChatGPT, Cursor, and Gemini integrations via MCP. Making OpenClaw a first-class OB1 citizen means every OpenClaw user gets multi-client memory for free, without OpenClaw having to build and maintain four different client integrations.

RAW_BUFFERClick to expand / collapse

Problem

OpenClaw's memory is session-local: MEMORY.md, memory/YYYY-MM-DD.md, and module CONTEXT.md files are brilliant for a single workspace but do not travel between clients. If I run an agent in a Claude Desktop session and a Cursor session, they each have their own memory of me.

Open Brain (OB1) by Nate B. Jones solves exactly this problem via an MCP server over Supabase/pgvector. Any MCP-capable AI client reads and writes the same thoughts table, so Claude, ChatGPT, Cursor, and Gemini all share a single brain.

Proposal

Add OpenBrain as a supported memory target for OpenClaw. Concretely:

  • A new config section memory.openbrain with url, accessKey, and mode (off | mirror | shared).
  • In mirror mode, tagged daily-memory entries are mirrored to OpenBrain on write. OpenBrain stays a read-only reference; canonical memory stays in files.
  • In shared mode, OpenClaw reads from both file memory and OpenBrain on memory_search, merged.
  • Ship it as an optional plugin. No OpenBrain dependency for users who do not want it.

What already exists that would feed this

I have a working reference implementation that does exactly the push / pull / search flow today, just outside the core. It handles:

  • MCP transport (JSON-RPC over streamable HTTP / SSE — the real shape OB1 returns).
  • Redaction pass for emails and common API keys before push.
  • Round-trip guard so you do not re-ingest your own captures.
  • Fingerprint-based idempotent re-runs via OB1's upsert_thought RPC.
  • Tag skip-list for noisy heartbeat-style entries.

It is MIT-licensed and I am happy to contribute the generalized version upstream.

Why this matters

OB1 already ships Claude Desktop, ChatGPT, Cursor, and Gemini integrations via MCP. Making OpenClaw a first-class OB1 citizen means every OpenClaw user gets multi-client memory for free, without OpenClaw having to build and maintain four different client integrations.

What I would like from maintainers

  1. Is this in scope? Or would you rather keep core memory strictly file-based and let users bolt on OB1 themselves?
  2. If in scope: design doc first, or a draft PR?

Happy to work on this either way.

— Simon

extent analysis

TL;DR

Integrate OpenBrain as a supported memory target for OpenClaw to enable shared memory across multiple clients.

Guidance

  • Review the existing reference implementation to understand the push/pull/search flow and its components, such as MCP transport and redaction pass.
  • Discuss the scope and design of the integration with maintainers to determine the best approach, including whether to keep core memory strictly file-based or support OpenBrain as an optional plugin.
  • Consider the benefits of integrating OpenBrain, including enabling multi-client memory for OpenClaw users without requiring separate client integrations.
  • Evaluate the proposed config section memory.openbrain and its modes (off, mirror, shared) to determine the best configuration options for users.

Example

No code example is provided as the issue does not contain specific code snippets that can be used as a reference.

Notes

The integration of OpenBrain with OpenClaw may require careful consideration of the design and implementation to ensure seamless interaction between the two systems. The existing reference implementation can serve as a starting point for the integration.

Recommendation

Apply workaround by integrating OpenBrain as an optional plugin to enable shared memory across multiple clients, allowing users to choose whether to use the feature or stick with the traditional file-based memory approach. This approach enables OpenClaw to leverage the benefits of OpenBrain without requiring a full overhaul of its core memory system.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…
#api#ssr#permission error#memory optimization#batch processing

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING

openclaw - 💡(How to fix) Fix Feature: OpenBrain (OB1) MCP as a first-class memory target [1 participants]