Making data portable. Cofounder @open_data_labs, creator of @vana. Prev: MIT, CSAIL research, janet yellen fan turned eth miner
San Francisco, CA
Joined April 2022
- Tweets800
- Following1.4K
- Followers23.9K
- Likes13.6K
Pinned Tweet
How data portability works on @vana:
- you control your data with your keys
- full audit trail of access, whether from agents or apps
- fast! ~50ms to permission data
Anna Kazlauskas retweeted
Sick of copying and pasting context between AIs?
V got you. That's what the Vana CLI is for.
Collect your data, run your own Personal Server, and let AI tools read from it. Only five commands to spin it up.
npm install -g vana-cli
vana login
vana connect github
vana connect chatgpt
claude mcp add vana -- vana mcp
Then your data goes wherever you go. Any AI you use can read your context, from any source you've connected - all from your Personal Server, on hardware you control and under keys only you hold.
Stop the server any time [vana server stop], and no app can reach your data until you start it again.
🔗👇
Best way to manage connecting a bunch of accounts to an agent is a @vana personal server:
- Use existing login details from your device (browser session stores etc)
- Choose where it's hosted (can be fully local)
- Connect it to any agent, see clear logs of what got accessed
Some early internet researchers thought owning your data meant physically wearing it. Instead data has ended up scattered all across platforms
Run a Vana personal server with all your private data and securely bring it across apps and agents (clunky hard drive belt optional)
The Vega upgrade is live, as set out in the January roadmap. Four things to know.
1. AI models are pretty good now. What sets one apart is context: what the model knows about the person asking. AI tools and agents look that up thousands of times a day. Every one of those lookups is your data moving. That is data portability.
2. On the internet, your data moves instantly. It moves between platforms without you in the loop. Vana moves your data where you choose and keeps your permission on record. The Vega upgrade brings that to the same speed: about a twentieth of a second.
3. Governments need the same thing. Most countries have laws that allow you to take your data with you. Almost none had a working way to do it until Vana wrote the standard for how personal data moves. It is now hosted by the Linux Foundation, and governments globally are already scoping national implementations.
4. The Vana network has always earned when pooled datasets were accessed. As of the Vega upgrade, it also earns every time one person's data moves. Every data portability transaction pays the network. Someone moves their data to an app. The app reads it. The network earns.
Want to know more? 🔗👇
Keep your memory yours
We've acquired the @MemoryProtocol team!
Vana has acquired the team behind @MemoryProtocol and dropped a major update to the Vana App.
every major AI added memory this year. five walled gardens. your context lives on their servers, serves their product, stops at their wall.
that's not very Vana. so, with the help of @JackSpallone, we've released something to keep you in control of your data
AI-driven hacks are exposing the absurdity of how personal data is stored
You technically own all your data, but right now platforms store it for you and try to protect it
You don't really own your data unless you know where it is and can delete it when you want
To prepare for the coming wave of AI, own your data:
- Store it locally
- Delete it from platforms you no longer use
- Port it across apps
All on @vana
Just to be clear.
The recommendation here is to own your data. Download it, store it locally, train your own models on it, let it learn about you, just don't give your identity away for free for "Better advertisements", which is just selling it to the highest bidder.
There are risks we don't yet understand. Be diligent. Live a life assuming everything will come to light.
Good luck. AGI is here. Even if it isn't broadly deployed.
If you run a data company then you also run a security company
Your ability to charge for your data depends on your ability to gate access to that data
As code becomes abundant, importance shifts to: data, provenance, protocols
@vana is this intersection
Good privacy-friendly workflow to bring your full chatGPT history (a very intimate dataset) into OpenClaw
- Doesn’t require waiting for GDPR export
- Keeps data local, assuming a local model
I told my @OpenClaw agent to pull my ChatGPT history in and realized it's probably the most intimate portrait of me that exists anywhere. Not surprised, considering it's thousands of conversations with a judgement-free non-human...
Engineering notes on a privacy-friendly way of pulling cross-platform data:
- import cookies from default browser
- make data connectors easily updatable by AI agents
- bundle a Vana personal server in to make data portable
There's a massive shift happening in software right now: AI is making platforms adversarial toward their own users. Platforms, from legacy software tools to new LLMs with built-in memory, want to keep your data and context locked in as their moat. But users want their context to travel with them across every AI application.
The legal framing is clear: you have full legal ownership of your data, the same way you own your car even if it's parked in a parking lot. But platforms can make it incredibly inconvenient to actually get it. GDPR and CCPA requests are buried deep in settings menus and take days to weeks to return your data.
Today we're releasing DataConnect, an open source tool that automatically grabs your deep, cross-platform data and stores it locally. It allows you to pull data that has never been available via API: ChatGPT memories, full Shopify purchase history, YouTube watch history, and makes it portable for any AI builder to use as context. Compute runs client-side, with data stored on your machine. You can build data connectors for any platform you want with a single Claude code prompt.
What would you build if a user showed up already carrying all their ChatGPT memories, their sleep history, and their music taste from the past few years?
Try it out: github.com/vana-com/data-con…
Builders want portable data, and users want portable data. The only groups that don't want portable data are those that are trying to build a moat around data they don't legally own.
Further reading on data portability:
- The Obsidian founder @kepano's before-its-time "file over app" philosophy on the importance of portable data: stephango.com/file-over-app
- A @scottbelsky piece on the coming “data wars” implications.com/i/161752760…
- A conversation with me, Representative @DougFiefia, and @sheila_warren on data portability and Utah's digital choice act: nitter.cf/i/broadcasts/1OyKALvEz…
Announcing DataConnect: an open source desktop app that pulls your cross-platform data in one click.
Platforms lock your context in: there's no API to get your Shopify purchases, ChatGPT memories + conversations, or Spotify listen history.
Try it out: account.vana.org/download-da…
ALT https://github.com/vana-com/data-connect
Anna Kazlauskas retweeted
Today is the day! Open Source AI Summit at @thehousesf from 3-7pm, cohosted with @NEARProtocol and sponsored by @gensynai @Celo @selfxyz @hashed_official 🫶🏻
luma.com/6ns88igh
We are expecting a full house, come by 2:30pm to ensure you are settled in before talks begin!💫
3:00 - 3:20pm @lukaszkaiser at OpenAI
Title: Open Source AI in the 2026 Age of Research
3:25 - 3:45pm @benfielding at @gensynai
Title: Prediction Markets are the Agentic Bazaar
3:50 - 4:10pm @ilblackdragon at @NEARProtocol
Title: Secure Your Personal AI Agent with IronClaw
4:15 - 4:30pm @RegenRene at @selfxyz
Title: Claw Control: Proof-of-Human Guardrails for Agentic Shenanigans
4:35 - 4:50pm @DavideCrapis @ethereumfndn
Title: SOS AI: Secure Open Source AI for Human Empowerment
4:55 - 5:10pm @marco_derossi at @MetaMask
Title: How Intelligence Grows in Agent Networks
5:15 - 5:40pm Fireside chat moderated by @andyyy at The Rollup, with @ilblackdragon at NEAR Protocol, @sreeramkannan at Eigenlayer
Title: Rise of Agentic Finance
5:40 - 5:55pm @Pierre_Le_Guen at @near_ai
Title: Beyond Mac Mini: Scaling Verifiable AI Agents with TEEs
6:00 - 6:15pm @JanLiphardt at @openmind_agi and @Stanford
Title: Do no harm - systems for human<>machine cooperation
6:20 - 6:35pm @annakaz at @vana
Title: Data Portability
6:40 - 7:05pm Fireside chat by @hosseeb at @dragonflyvc, @ai at Canonical x Lightspeed, @baekkyoumkim at @hashed_official
Title: Investor AI jam session
See you soon SF!
Interesting the Pope is arguing for more open/distributed control of AI
Unusual for builders to be aligned with and debating the same issues as the Pope
What a time to be alive
Human beings are called to be co-workers in the work of creation, not merely passive consumers of content generated by artificial technology. Our dignity lies in our ability to reflect, choose freely, love unconditionally, and enter into authentic relationships with others. Recognizing and safeguarding what characterizes the human person and guarantees their balanced growth is essential for establishing an adequate framework to manage the consequences of artificial intelligence.
vatican.va/content/leo-xiv/e…
Messaging apps like Signal/Telegram pioneered end-to-end encryption, which led to mainstream players like WhatsApp adopting stronger privacy guarantees
There's potential for something similar to play out in confidential computing: early adopters will shift consumer expectations to have their data handled securely (and not used to train AI without their consent)
Technical details: cocoon.org/docs/tdx-and-imag…
🐣 It happened. Our decentralized confidential compute network, Cocoon, is live. The first AI requests from users are now being processed by Cocoon with 100% confidentiality. GPU owners are already earning TON. cocoon.org is up.
🏦 Centralized compute providers such as Amazon and Microsoft act as expensive intermediaries that drive up prices and reduce privacy. Cocoon solves both the economic and confidentiality issues associated with legacy AI compute providers.
📈 Now we scale. Over the next few weeks, we’ll be onboarding more GPU supply and bringing in more developer demand to Cocoon. Telegram users can expect new AI-related features built on 100% confidentiality. Cocoon will bring control and privacy back where they belong — with users.
Anna Kazlauskas retweeted
1/ Grain, oil, and equity began as messy local exchanges before standards were set. We’ve yet to see this for training data.
Today’s data deals remain opaque and ad hoc, from News Corp with OpenAI ($250M+), to Reddit with Google ($60M per year), to HarperCollins with Microsoft ($5,000 per book).
In a new paper with @Annakaz for @vana & @open_data_labs, we chart a path forward.
Full paper: arxiv.org/abs/2510.24990
