Eilon Mor

CTO & Co-Founder at

[Eilon Mor @MorEilon]

@garrytan Nice, I'm going to try it. How are the meeting transcripts automatically getting into the gbrain repo? Is there a native integration with data sources like note-takers, calendar, etc., or should those be integrated separately into the OpenClaw machine?

[Eilon Mor @MorEilon]

@DrewAustin https://t.co/9VZgEAQFFD

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

We were obsessed with the same problem.

Eventually we went with a different approach, building on a similar concept to mempalace (https://t.co/dUmRbsymC0). Our meeting transcripts automatically map and build structured knowledge, which then auto-feeds a team of OpenClaw agents.

Just posted a short demo.

[Eilon Mor @MorEilon]

@AlexFinn https://t.co/9VZgEAQFFD

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

@AlexFinn i went a step further and merged it all into one system. meetings get captured and transcribed automatically, flow into an obsidian-style knowledge graph, and OpenClaw agent teams pull from it. meetings go in, connected knowledge comes out, and OpenClaw agents start working.

[Eilon Mor @MorEilon]

@Victor_Webflow Just shipped this: a knowledge graph that builds itself automatically from your meetings. People, decisions, problems, projects, all auto-linked.

https://t.co/9VZgEAQFFD

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

@CiprianiRanieri https://t.co/9VZgEAQFFD

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

@audiencon Built a memory system that turns past meetings into a knowledge graph. People, decisions, projects, all auto-connected. No tagging, no curation.

https://t.co/9VZgEAQFFD

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

Early Access → https://t.co/ecSPDBfedS

[Eilon Mor @MorEilon]

What if your knowledge base didn't need you at all?

I built a memory system that turns your past meetings into a knowledge graph. People, decisions, problems, projects, vendors, auto-connected. No tagging. No curation. It just builds.

AI agents now remember what your team forgot.

[Eilon Mor @MorEilon]

@AlexFinn We found all the Anthropic detection triggers - system prompt keywords, tool names, user-agent, billing headers. All string matches. Full breakdown + fix here:

https://t.co/3Mgw7xtgrX

[Eilon Mor @MorEilon]

Spent a day reverse-engineering how Anthropic detects OpenClaw agents on setup tokens. Found multiple detection layers. All fixable with a proxy.

Here's every trigger and how to bypass them 👇

[Eilon Mor @MorEilon]

@NotSoEasyMoney We reverse-engineered every detection layer. All string matches are easily bypassable with a proxy. Full breakdown:

https://t.co/3Mgw7xtgrX

[Eilon Mor @MorEilon]

Spent a day reverse-engineering how Anthropic detects OpenClaw agents on setup tokens. Found multiple detection layers. All fixable with a proxy.

Here's every trigger and how to bypass them 👇

[Eilon Mor @MorEilon]

@sanjaybuilds_ There's a third option - an LLM proxy that bypasses the detection. We reverse-engineered every trigger.

Full breakdown:
https://t.co/3Mgw7xtgrX

[Eilon Mor @MorEilon]

Spent a day reverse-engineering how Anthropic detects OpenClaw agents on setup tokens. Found multiple detection layers. All fixable with a proxy.

Here's every trigger and how to bypass them 👇

[Eilon Mor @MorEilon]

@Voxyz_ai We found all the triggers - system prompt keywords, tool names, user-agent, billing headers. All string matches, no ML. Full breakdown + fix here:

https://t.co/3Mgw7xtgrX

[Eilon Mor @MorEilon]

Spent a day reverse-engineering how Anthropic detects OpenClaw agents on setup tokens. Found multiple detection layers. All fixable with a proxy.

Here's every trigger and how to bypass them 👇

[Eilon Mor @MorEilon]

Binary searched the 95KB request body to find each trigger. All fixable with string replacements in the proxy.

We handle this automatically if you use setup tokens on https://t.co/J5yBSuIQ1f

[Eilon Mor @MorEilon]

Spent a day reverse-engineering how Anthropic detects OpenClaw agents on setup tokens. Found multiple detection layers. All fixable with a proxy.

Here's every trigger and how to bypass them 👇

[Peter Steinberger 🦞 @steipete]

Anthropic now blocks first-party harness use too 👀

claude -p --append-system-prompt 'A personal assistant running inside OpenClaw.' 'is clawd here?'

→ 400 Third-party apps now draw from your extra usage, not your plan limits.

So yeah: bring your own coin 🪙🦞

[Eilon Mor @MorEilon]

→ Rename sessions_* tool names outbound, reverse-rename in the response stream