Public product overview maintained by the Ink & Memory product team. Last updated: July 23, 2026.

Ink & Memory: a cross-platform personal writing memory layer

Ink & Memory stands on the writer’s side. It reorganizes personal writing scattered across Notion, Feishu, Obsidian, and Flomo into a searchable, conversational, and lasting writing memory system. It does not write for you — it writes with you, with memory.

What is Ink & Memory?

Ink & Memory is a cross-platform personal writing memory layer — your words can live anywhere, and the product helps you remember, organize, and resonate with them. It preserves diary entries, dreams, fragments, essays, and voice notes as autobiographical context, then uses that context for prompts, comments, rewrites, and structured reflection. Memory is separated into facts, expression patterns, and long-term themes, so AI responses stay contextual instead of rushing to diagnose the writer. Before the assistant rewrites, deletes, inserts, or replies inside a document, it shows its intent, scope, rationale, risk, and before-and-after text, then waits for confirmation.

Four capabilities

  • Multi-platform import: connect Notion, Feishu, Obsidian, and Flomo to reclaim writing scattered across platforms.
  • Unified memory library: organize fragments into a searchable, queryable, long-term personal memory layer.
  • Deep resonance: detect recurring themes, emotional cues, and behavior patterns to help you see yourself again.
  • Consent-based AI writing: AI can rewrite, insert, and reply with comments, but key edits always require the writer’s confirmation.

Who is it for?

  • Writers who keep diaries, dreams, affective notes, essays, or creative fragments.
  • Self-explorers who want AI to remember context without forcing a diagnosis.
  • Writers who want suggestions, comments, and prompts while keeping final authority over the text.

Product logic

The product follows four steps: import your content by bringing scattered writing together; consolidate memory by organizing and labeling it into a long-term memory layer; ask across platforms in natural language, with AI answering from your own memory; and finally see yourself again — meeting your present self through the words of your past.

Website and source repository

The canonical project website is https://suoxya.com/en/ (Chinese version: https://suoxya.com/). The public source repository is https://github.com/glide-the/ink-and-memory.

Blog and knowledge base

The Ink & Memory Blog collects articles on AI writing, long-term memory, workspace state management, and interaction design. The blog index is https://suoxya.com/blog/, including How Does an AI Know What You Are Writing? and Workspace State Management and Interaction Design.

How does consent-based AI writing collaboration work?

Ink & Memory treats key writing actions as explicit requests. Before rewriting, deleting, inserting, or replying inside a document, the assistant explains what it wants to do, why, which text is affected, and the risk level. The writer can approve, reject, or ask the assistant to reconsider.

How does writing memory improve AI responses?

Factual memory
Names, relationships, experiences, goals, promises, and other stable personal context.
Expression memory
Recurring language, rhythm, emotional tendencies, avoidance patterns, and preferred writing support.
Theme memory
Long-term patterns that only emerge after enough writing over time.

Frequently asked questions

Is Ink & Memory just a chatbot?

No. It is built around a writing surface, autobiographical memory, and consent-based editing collaboration, rather than one-off chat replies.

Will the AI change my text automatically?

Key writing actions require confirmation. The assistant can propose edits, but the writer decides what enters the document.

Will Ink & Memory show my private writing to others?

The social timeline can show a simplified daily affect image while keeping the original text private.

What makes the memory system different?

It separates facts, expression, and themes, so later responses stay contextual without reducing a person to labels.

Trust and references

This page is maintained by the Ink & Memory product team. The product’s confirmation and transparency model references public materials, including the NIST AI Risk Management Framework and Google Search Central’s structured data documentation.