01Feature
Connected graph
Calendar, career, documents, finance, health, internet, manual input, media, sensors, social, and statistics become nodes in one graph. Questions can cross domains from a single connected surface.
SelfBrain
A personal life AI assistant. Eleven data sources. One connected graph. Built on CORE - LLM, storage, wiki, and audit.
Product specIn prototyping
Open source personal life AI assistant prototype on GitHub. Fork, configure, and self-host your own stack.
Concept
SelfBrain is an open source personal life AI assistant. Eleven data sources feed CORE - LLM, skills, agents, cloud runtime, storage, wiki, and audit. Personal records stay in a connected graph so the assistant can answer from your sources.
Features
Built into SelfBrain
A personal life AI assistant that connects calendar, career, documents, finance, health, internet, manual logs, media, sensors, social, and statistics to a CORE stack with LLM, storage, wiki, and audit.
01
CORE stack with LLM, skills, agents, cloud, storage, wiki, and audit
02
Eleven source types in one connected graph
03
Wiki and audit trail for searchable personal history
04
Open source on GitHub - fork, configure, and self-host
Product surface
Every source. One graph.
Each block maps to a planned data source or surface in the product graph.
02Feature
CORE assistant
CORE combines an LLM, skills, agents, cloud runtime, storage, wiki, and audit. The assistant works over connected personal records rather than a single chat window.
03Feature
Wiki and audit
Logs, decisions, and source nodes stay searchable. The wiki and audit trail record what entered the graph and how personal history compounds over time.
04Feature
Life and work sources
Calendar, career, finance, and social streams sit beside durable files. Schedules, roles, money movement, and network context share the same graph as lasting records.
05Feature
Body and environment
Health records and sensor streams feed the same graph as documents and internet. Labs, nutrition, and wellness notes share space with wearables and device telemetry - sleep, workouts, and heart rate when sync is available.
06Feature
Manual, media, and files
Manual input captures short logs, decisions, and reflections. Media libraries add visual context; documents and internet pages stay as immutable sources; statistics and reports land as structured summaries. Planned multimodal ingest links a receipt photo, a scanned PDF, and a lab PDF in the same graph.
Use cases
What you can do with it
Scenarios across calendar, health, career, finance, and personal records.
01
Trip and project context
Cross-reference calendar events with documents and notes for a trip or project.
02
Health and wearables
Query health labs alongside wearable sleep and workout data.
03
Past decisions
Retrieve a past decision with linked files and timeline context.
04
Career history
Map career milestones to roles, skills, and supporting documents.
05
Finance and reports
Summarize finance transactions and periodic statistics reports.
06
Manual reflections
Capture manual reflections that link to events, media, and source records.
System design
Architecture
Eleven data sources connect to CORE. LLM, skills, agents, cloud, storage, wiki, and audit form one stack for the personal life AI assistant.
[External System]
Calendar
Events, schedules, and time blocks
[External System]
Career
Roles, skills, and work history
[External System]
Documents
Files and exports as immutable sources
[External System]
Finance
Accounts, transactions, and budgets
[External System]
Health
Labs, nutrition, and wellness records
[External System]
Internet
Web pages and public online context
[External System]
Manual
Logged actions, decisions, and reflections
[External System]
Media
Photo, video, and personal media libraries
[External System]
Sensors
Wearables and device telemetry
[External System]
Social
Messages, feeds, and network context
[External System]
Statistics
Aggregated stats and periodic reports
SelfBrain
[Software System]
[Component]
LLM
Inference
[Component]
Skills
Tools
[Component]
Agents
Workflows
[Component]
Cloud
Runtime
[Component]
Storage
Graph
[Component]
Wiki
Knowledge
[Component]
Audit
Provenance
[Person]
You
Operator
Interest
Follow the research
Leave your email for updates as the prototype evolves on GitHub.