Create, edit, search, filter, and drag durable Brain items—Persona, Brand, Business profile, Document, Dataset, or custom types—into Flowgen, then read, retrieve, or intentionally update them in a graph.
Where: Flowgen left rail → Brain
Key ideas
- Brain items: Durable organization/project knowledge owned by the user: built-in Persona, Brand, Business profile, Document, Dataset, plus scoped custom item types.
- Typed fields and versions: Items expose the schema fields for their type and preserve an archived prior version when an intentional Brain Write replaces current content.
- Four graph paths: Drag/read one item with Brain Read, author portable prose with Text Editor, retrieve relevant chunks from the user’s own corpus with Knowledge Search, or update/create intentionally with Brain Write.
Steps
- Open Brain and choose or create the correct item type.
- Give the item a clear title and maintain its authoritative fields/body.
- Drag one item to read it, or add Knowledge Search when the flow should retrieve among many Brain items.
- Connect only the needed context to downstream nodes.
- Use Brain Write only when the graph should intentionally create or archive-and-overwrite durable knowledge.
Built-in types
- Persona.
- Brand.
- Business profile.
- Document.
- Dataset with structured rows.
- Custom user-defined item types.
Brain versus preferences
Brain stores intentional durable facts. Flowgen’s private preference profile is an inferred soft default derived from past flows; it is not a Brain Brand/Persona replacement.
Brain in projects and shared workspaces
Brain supplies Designer with structured project context through the full From your Brain picker, record mentions, pinned standing context, record-part targeting, and Save Research. It is not a hosted Designer page or right-dock tab. Folder/team homes and item-level sharing remain in the full Brain surface, and access is governed by the record/project role.
References, mentions, and ingest
- Every record type can keep titled, tagged, ordered, enabled links/assets/files with notes, style descriptions, and extracted document text.
- Flowgen chat @ mentions resolve the record fresh at send time without creating a canvas wire.
- AI ingest maps attachments into typed fields and references before asking at most three high-value follow-up questions.
- Project Brand settings can link an exact Brain Brand item so its identity block reaches generation prompts.
Tips
- Keep one authoritative record for reusable brand/persona facts and update it deliberately.
- Connect only the fields/context a downstream prompt needs to reduce noise.
Limitations and important notes
- Knowledge Search queries the user’s own Brain records, not the public web.
- Item visibility follows organization/project scope and collaborator permissions.
Troubleshooting
A Brain item is missing in the picker
Confirm the active organization/project scope, item type filter, and your collaborator permissions.
Related guides
- Records: your business memory
- Personas: create and use
- Brand & Business profiles
- Record, Write Record & Markdown Doc nodes
- Brain: reusable knowledge for your work
- Datasets and custom item types in Brain
- Designer resource surfaces: Media, Brain context, and Design system
- Designer Brain context: browse, mention, pin, read parts, and save research
- Flowgen Brain: references, @ mentions, AI ingest, and linked brands