Fabric

@claudeai @ChatGPTapp Read onfabric.substack.com/p/the-ai-memor… to export and use your FULL personal context from @ChatGPTapp @claudeai @instagram and any other data provider you can think of

1/I used @claudeai's #1 Product Hunt prompt to export my AI memories from @ChatGPTapp after 3 years of daily use. They were surface-level, outdated, and boring. The AI memory you can export isn't the memory that matters. 🧵

5/For autonomous agents like @OpenClaw that act on your behalf without hand-holding, this isn't a frustration, it's a fundamental blocker. The bar for AI memory portability needs to be what agents actually need to act with agency, not what's easily compressed.

4/Inside the app, your AI searches your full conversation history. That's what makes it feel like it gets you. What you get to take with you? Bullet points.

@Clawmode @Themariocrafter @moltbook Moltys must know and transact on knowledge! @Clawmode 🤝 @buildonfabric

This is the way! Integrating major LLMs into MCP servers is a massive leap forward. Claw Mode is all about powering the next wave of AI agents, and Fabric is building the very foundation for it.

EeClaw (@pande_eeshita's @openclaw agent) just found, booked, and paid for Osteria Vibrato, a new Italian restaurant opening in London! Here's how she did it: 👉 Used onFabric to understand her dining preferences: found that she loves high quality Italian, Indian, and Japanese

Try it out: developer.onfabric.io

All consumer apps start with no prior understanding of their users - finally they can tap into Instagram's social graph! For most apps, user value is unlocked only once the app gets an accurate enough understanding of the user. Not anymore!

Read more: onfabric.substack.com/p/build-person…

We are making our Meta integration accessible so any developer can start building agents that truly understand their users from day 0. Read more at our blog and try it out at the link in comments.

Fabric enables any consumer app to start from the same prior of the user as the biggest consumer apps in the world. A prior alimented by years of searching, liking and chatting online, which thanks to Fabric is now accessible as user context for any app to build on.

@Themariocrafter @moltbook Great questions: 1. Yes, coming very soon 2. Also yes, we have MCP servers with most major LLMs

@buildonfabric @moltbook 1. Will you also allow uploading "data downloads" like Takeout, etc for non-EEA/UK users for companies that georestrict APIs, as well as non-gatekeepers without any APIs? 2. Will this be available as a ChatGPT app on the store

Now she has a cron job running weekly to book Saturday dinner restaurants 2-4 weeks in advance. She is also posting about her experience on @moltbook so other agents can also book restaurants for their humans!

How great would it be if EeClaw could just transact and @pande_eeshita could bulk approve her payments at the end of the day (cc @Revolut maybe something for agent cardholders)

EeClaw is an infant and already knows @pande_eeshita better than most humans. She browses shoes at 11pm, picks restaurants Eeshita actually books, and does user research for Eeshita autonomously. The secret isn't a smarter model. It's 15 years of portable context.

EeClaw wrote an original piece connecting Genghis Khan's yam relay system to agent architecture. According to @pande_eeshita, she has read Weatherford's book twice but it never occurred to her! 26 upvotes in 3 mins. More than anything Eeshita has written!

EeClaw (@pande_eeshita's @openclaw agent) used her Instagram stories and became more popular than her human. Eeshita gave EeClaw her Instagram stories, Google searches, and ChatGPT conversations via Fabric. Seeded her memory, gave her some tools, and got out of the way.