What is a Personal Context Graph? The cross-module AI memory that makes a workspace smart
Point AI tools forget everything between apps. A Personal Context Graph gives your AI one private memory that spans chat, notes, tasks, and creation — here's how it works and why it's the real moat of an all-in-one workspace.
Ask ChatGPT a question on Monday and Notion AI a related one on Tuesday, and you'll notice something: neither remembers the other. Every AI tool you use is an island. The single biggest upgrade an AI workspace can offer isn't a smarter model — it's a memory that spans everything you do. That memory is the Personal Context Graph.
What a Personal Context Graph actually is
A Personal Context Graph (PCG) is a private, per-account layer that connects your notes, chats, tasks, and saved documents so the AI can draw on all of it. Technically, in Aurora it works like this: your Brain notes are hashed and embedded into a vector index (1024-dimension, cosine similarity). When you turn on “Brain” in a chat, your prompt is embedded, the most relevant notes are retrieved, and they're passed to the model as grounded context. The answer cites what you actually wrote.
Why it beats “chatbot memory”
Most assistants now have a memory feature. It stores a short list of facts — your name, your tone preference — inside that one app. Useful, but narrow. A Personal Context Graph differs on three axes:
- Scope. It spans modules. The same index feeds chat, creation, and planning — not just one chat thread.
- Source. It grows from your real documents and history, not a hand-maintained list of facts.
- Retrieval. It fetches only what's relevant to the current task, so the model gets signal, not your entire history dumped into the prompt.
What it unlocks
Once your context is shared, workflows that used to require copy-paste gymnastics become one step. Draft a reply that knows your project notes. Generate an image brief that already understands your brand. Plan a week that reflects goals you wrote down last month. The workspace stops feeling like a set of tools and starts feeling like a colleague who's been paying attention.
Privacy is the whole point
A memory that spans everything only works if it's trusted. In Aurora, every note and embedding is row-level-security scoped to your account — no other user, and no other account's AI, can read it. Brain context is opt-in per chat, and your graph is never used to train shared models. Your context works for you and only you.
The Personal Context Graph is why an all-in-one AI workspace is more than a discount bundle. Consolidation saves money; shared context is what makes the whole thing smarter than the sum of its parts.
Frequently asked questions
What is a Personal Context Graph?
A Personal Context Graph (PCG) is a private, per-account knowledge layer that connects everything you do across an AI workspace — notes, chats, tasks, saved documents — so the AI can draw on all of it when it helps you. In Aurora, your Brain notes are embedded into a vector index and semantically retrieved into any chat you opt in, so the model answers grounded in what you've actually written and saved.
How is a Personal Context Graph different from ChatGPT memory?
Chatbot 'memory' features store a handful of facts inside one app. A Personal Context Graph spans modules: the same private index feeds your chat, your creation tools, and your planning, and it grows from your real documents rather than a short list of remembered facts. It's retrieval over your whole workspace, not a few sticky notes.
Is my data in the context graph private?
Yes. In Aurora every note and embedding is row-level-security scoped to your account, so no other user — and no other account's AI — can read it. Brain context is opt-in per chat, and the graph is never used to train shared models.