Onelist Ecosystem Strategy - The Memory Layer for AI Tools

Capability Matrix

Click any row to expand for details on what each capability means.

Capability	Onelist	Twin	Supermemory	OpenClaw
Memory Storage	✓ Full	◐ Basic	✓ Full	◐ Basic
Memory Storage refers to the ability to persistently store and organize knowledge, documents, and memories. Onelist provides full PostgreSQL-backed storage with versioning, tagging, and hierarchical organization. Twin and OpenClaw have basic session/context memory but aren't designed as primary knowledge stores.
Atomic Extraction	✓ Core	✗	✓ Core	✗
Atomic Extraction breaks long-form content into discrete, searchable facts with resolved references (e.g., "Sarah mentioned the budget is $50K" becomes a standalone memory linked to its source). Onelist's Reader Agent performs this extraction with reference resolution, similar to Supermemory's approach. This enables precise retrieval of specific facts rather than just matching documents.
Semantic Search	✓ Two-layer	◐	✓ SOTA	◐
Semantic Search finds content by meaning, not just keywords. Onelist implements two-layer retrieval: search on atomic memories for precision, then return source chunks for full context. This combines the best of both approaches - precise fact-finding with rich context. Includes similarity checking to prevent duplicates and temporal metadata for time-aware queries.
Knowledge Organization	✓ GTD	✗	✗	✗
Knowledge Organization is a methodology for structuring information. Onelist implements full GTD (Getting Things Done) with inbox, next actions, waiting for, someday/maybe buckets, contexts (@phone, @computer), projects, areas, and weekly reviews. River Agent guides users through this system. Other platforms lack structured organizational methodology.
User Interface	✓ Full Stack	✓ Dashboard	✗ API only	◐ Chat
User Interface refers to how you interact with the platform. Onelist provides the full stack: Web app for browser access, Desktop app for native experience, Chat via River Agent for conversational interaction, and REST API for programmatic access and integrations. Twin has a workflow dashboard. Supermemory is API-only (developers build their own UI). OpenClaw is primarily chat-based through messaging platforms.
Asset Enrichment	✓ Full	✗	✗	✗
Asset Enrichment automatically transforms raw uploads into searchable, actionable knowledge. Onelist's Asset Enrichment Agent processes: Images: AI descriptions, OCR text extraction, face recognition linked to person entries - search "my daughter's first steps" and find that video by connecting faces, birthdays, age, and enrichment metadata Audio: Whisper transcription with timestamps, speaker diarization, automatic action item extraction from meetings Video: Keyframe analysis, audio extraction, scene detection, chapter generation Documents: PDF text extraction, table parsing, structure understanding Extracted action items and decisions become first-class entries linked to their source. Other platforms don't offer built-in asset intelligence.
Content Capture	✓ Feeder Agent	✓ Scraping	✓ Connectors	✓ CDP
Content Capture is how content gets into the system. Onelist's Feeder Agent is the unified gateway for all external content: Web Clipper (MVP): Browser extension for URL capture, article extraction (Readability-style), YouTube transcripts, with duplicate detection via Searcher Agent One-time Import: Bulk import from Evernote (ENEX), Notion (ZIP), Obsidian vaults, Apple Notes exports Continuous Sync: RSS feeds, Evernote API, Notion OAuth, Obsidian REST API, Lifelog Audio, Email (IMAP), Instapaper, Twitter/X Bookmarks Cloud Storage (Post-MVP): Google Drive, Dropbox, OneDrive integration - files stay in your cloud storage, Onelist indexes and enriches them without duplication Twin scrapes via browser automation. Supermemory has connectors for Google Drive, Notion, OneDrive, S3. OpenClaw uses Chrome DevTools Protocol.
Task Automation	✓ Tools + MCP	✓ Browser	✗	✓ Shell + CDP
Task Automation means executing actions in external systems. River Agent supports three handler types: Internal tools: GTD operations, search, entries, briefings, notifications - all core Onelist operations Webhook tools: HTTP POST to any external service - user-defined integrations with any API MCP tools: Model Context Protocol integrations - calendar, email, and other MCP-compatible services Skills System (Post-MVP): OpenClaw-style text-based skill instructions for complex workflows Twin focuses on browser-based automation (2,700+ apps). OpenClaw executes shell commands and uses Chrome DevTools Protocol (CDP) for browser control.
Messaging Integration	◐ Planned	✗	✗	✓ 12+
Messaging Integration connects to chat platforms like Telegram, WhatsApp, Discord, and Slack. OpenClaw is the leader here with 12+ messaging integrations. Onelist plans to add external channel support (Telegram, Slack) for River Agent interactions post-MVP. Twin and Supermemory don't focus on messaging.
Self-Hosted	✓ Free	✗	✗	✓ Core
Self-Hosted means you can run the software on your own infrastructure with complete data ownership. Onelist and OpenClaw are both self-hosted first, free forever to run on your own servers. Twin and Supermemory are cloud-only SaaS products with no self-hosting option.
E2E Encryption	✓	✗	✗	✗
E2E Encryption means your data is encrypted before leaving your device, so even the cloud provider cannot read it. Onelist's Cloud Sync tier uses end-to-end encryption - we literally cannot access your data. Other platforms either don't offer cloud sync (OpenClaw) or process data on their servers (Twin, Supermemory).
MCP Server	◐ Planned	✗	✓	✗
MCP Server (Model Context Protocol) enables portable memory across AI tools - use the same knowledge base with Claude, ChatGPT, or any MCP-compatible assistant. Supermemory pioneered this with their MCP server. Onelist plans to expose Searcher Agent as an MCP server post-MVP, enabling memory portability across tools.
Open Source	✓ MIT	✗	◐ Partial	✓ MIT
Open Source means the code is publicly available to audit, modify, and self-host. Onelist and OpenClaw are both fully open source under the MIT license - no vendor lock-in, full transparency. Supermemory has open-sourced their benchmarks but the core platform is proprietary. Twin is entirely proprietary.

✓ = Full support ◐ = Partial/Planned ✗ = Not available

🤖

Twin Integration

Onelist as the privacy layer for business automation

What Twin Does Well

✓ Browser automation with computer vision
✓ 2,700+ pre-built app integrations
✓ Zero-setup agent creation in plain English
✓ Self-correcting agents that adapt

What Twin Users Need

→
"My agents forget context"
Onelist provides persistent atomic memories across sessions
→
"I need self-hosted for compliance"
Onelist + Twin API = privacy-compliant automation
→
"Where's all the data Twin extracted?"
GTD-organized entry storage with relationships

Integration Architecture

┌─────────────────────────┐         ┌─────────────────────────┐
│         TWIN            │         │        ONELIST          │
│                         │  REST   │                         │
│  ┌─────────────────┐    │   API   │  ┌─────────────────┐    │
│  │  Browser Agent  │────┼────────►│  │  Reader Agent   │    │
│  │  (scrape data)  │    │         │  │ (extract memory)│    │
│  └─────────────────┘    │         │  └─────────────────┘    │
│                         │         │           │             │
│  ┌─────────────────┐    │         │           ▼             │
│  │  Workflow Agent │────┼────────►│  ┌─────────────────┐    │
│  │  (needs context)│◄───┼─────────┤  │ Searcher Agent  │    │
│  └─────────────────┘    │         │  │ (provide context│    │
│                         │         │  └─────────────────┘    │
│  ┌─────────────────┐    │         │           │             │
│  │  2,700+ Apps    │    │         │           ▼             │
│  └─────────────────┘    │         │  ┌─────────────────┐    │
│                         │         │  │  Entry Storage  │    │
└─────────────────────────┘         │  │ (E2E encrypted) │    │
                                    │  └─────────────────┘    │
                                    └─────────────────────────┘

Implementation: Twin Workflow Using Onelist

# Twin workflow using Onelist for memory
name: Research with Memory
trigger: manual

steps:
  # Step 1: Get context from Onelist
  - action: http_request
    config:
      url: "https://your-onelist.com/api/v1/search"
      method: POST
      headers:
        Authorization: "Bearer ${ONELIST_API_KEY}"
      body:
        query: ""
        limit: 10
    output: existing_knowledge

  # Step 2: Twin does browser research
  - action: browser_scrape
    config:
      urls: [""]
      extract: ["title", "content", "date"]
    output: new_research

  # Step 3: Store findings in Onelist
  - action: http_request
    config:
      url: "https://your-onelist.com/api/v1/entries"
      method: POST
      body:
        title: "Research: "
        content: ""
        tags: ["research", "twin-automated"]

View full Twin comparison

🧠

Supermemory Integration

Complementary positioning in the memory space

What Supermemory Does Well

✓ 81.6% LongMemEval benchmark (SOTA)
✓ Sub-300ms retrieval latency
✓ MCP server for AI tool integration
✓ Intelligent forgetting & temporal reasoning

Where Onelist Complements

→
"I need a UI for my memory"
Onelist provides complete web/desktop interface
→
"Users want to self-host"
Onelist self-hosted with Supermemory-class memory
→
"Need productivity methodology"
River Agent, Weekly Review, Six Horizons of Focus

Market Segmentation

Supermemory Wins

• AI developers building apps (API-first)
• Maximum retrieval accuracy needs
• Enterprise with proven scale requirements

Onelist Wins

• Privacy-conscious users (self-hosted)
• End users wanting UI, not just API
• GTD/productivity methodology users
• Enterprise with compliance requirements

MCP Interoperability

Both can expose MCP servers, letting AI assistants choose their memory backend:

// User chooses which backend
const memoryBackend = process.env.MEMORY_BACKEND;

const mcpConfig = {
  servers: {
    memory: memoryBackend === "onelist"
      ? "http://localhost:4000/mcp"  // Self-hosted
      : "https://api.supermemory.ai/mcp"  // Cloud
  }
};

View full Supermemory comparison

🦞

OpenClaw Integration

The perfect partnership - OpenClaw does, Onelist remembers

"OpenClaw does. Onelist remembers."

OpenClaw excels at executing tasks. Onelist excels at storing knowledge. Together, they create the complete AI assistant.

What OpenClaw Does Well

✓ 12+ messaging platforms (WhatsApp, Telegram, etc.)
✓ Browser control via Chrome DevTools Protocol
✓ Proactive behavior - reaches out first
✓ 100K+ GitHub stars, massive community

What OpenClaw Users Need

→
"My AI forgets conversations"
Atomic memories that never forget
→
"Can't find past information"
Two-layer retrieval, GTD organization
→
"Security concerns"
E2E encryption, read-mostly architecture

Integration Architecture

                           USER
                             │
                 WhatsApp / Telegram / Discord
                             │
                             ▼
┌─────────────────────────────────────────────────────────┐
│                      OPENCLAW                          │
│                                                         │
│  ┌──────────────┐  ┌──────────────┐  ┌──────────────┐  │
│  │   Gateway    │  │   Browser    │  │    Shell     │  │
│  │  (messages)  │  │    (CDP)     │  │  (commands)  │  │
│  └──────────────┘  └──────────────┘  └──────────────┘  │
│          │                                              │
│          │  "What do I know about Project X?"          │
│          ▼                                              │
│  ┌──────────────────────────────────────────────────┐  │
│  │           ONELIST SKILL (ClawdHub)               │  │
│  │  • remember  • recall  • inbox  • review         │  │
│  └──────────────────────────────────────────────────┘  │
│                          │ MCP / REST                   │
└──────────────────────────┼──────────────────────────────┘
                           ▼
┌─────────────────────────────────────────────────────────┐
│                       ONELIST                          │
│                                                         │
│  ┌────────────┐  ┌────────────┐  ┌────────────┐        │
│  │   Reader   │  │  Searcher  │  │   River    │        │
│  │  (extract) │  │   (find)   │  │   (GTD)    │        │
│  └────────────┘  └────────────┘  └────────────┘        │
│                         │                               │
│                         ▼                               │
│  ┌──────────────────────────────────────────────────┐  │
│  │           Entry Storage (E2E encrypted)          │  │
│  └──────────────────────────────────────────────────┘  │
└─────────────────────────────────────────────────────────┘

ClawdHub Skill Implementation

# SKILL.md for ClawdHub
---
name: onelist
version: 1.0.0
author: onelist
description: Give OpenClaw perfect memory
tags: [memory, knowledge, gtd, productivity]
requires:
  - ONELIST_API_URL
  - ONELIST_API_KEY
---

# Commands

remember <content>  # Store in Onelist
recall <query>      # Search memories
inbox <item>        # Add to GTD inbox
review              # Daily/weekly review

Quick Install

openclaw skill install onelist

Example Usage

"remember that John's birthday is March 15th"

"What do I recall about the Q3 budget?"

"inbox call mom about weekend plans"

"Show me my review for this week"

User Experience Example

👤

You (via WhatsApp):

What did we decide about the marketing budget?

🦞

OpenClaw → Onelist:

/memory/search?q="marketing budget decision"

Onelist returns:

Q1 budget: $450K (up 15%)
Digital spend: 60% ($270K)
Events: $80K for 3 conferences

🦞

OpenClaw → You:

Based on your notes from December 15th: Q1 marketing budget is $450K, with 60% ($270K) for digital and $80K for events. Want me to pull up the full document?

View full OpenClaw comparison

Onelist as the Universal Memory Layer

┌─────────────────────────────────────────────────────────────────────────────┐
│                       AI TOOL ECOSYSTEM                                      │
├─────────────────────────────────────────────────────────────────────────────┤
│                                                                              │
│  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌──────────┐  ┌──────────┐      │
│  │   Twin   │  │ OpenClaw │  │  Claude  │  │ ChatGPT  │  │  Custom  │      │
│  │          │  │          │  │   Code   │  │          │  │   Apps   │      │
│  └────┬─────┘  └────┬─────┘  └────┬─────┘  └────┬─────┘  └────┬─────┘      │
│       │             │             │             │             │             │
│       └─────────────┴─────────────┴─────────────┴─────────────┘             │
│                                   │                                          │
│                         MCP / REST API / SDK                                 │
│                                   │                                          │
│                                   ▼                                          │
│  ╔═══════════════════════════════════════════════════════════════════════╗  │
│  ║                           ONELIST                                      ║  │
│  ║                                                                        ║  │
│  ║   ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  ┌─────────────┐  ║  │
│  ║   │   Atomic    │  │  Two-Layer  │  │     GTD     │  │     E2E     │  ║  │
│  ║   │  Memories   │  │  Retrieval  │  │Organization │  │  Encryption │  ║  │
│  ║   └─────────────┘  └─────────────┘  └─────────────┘  └─────────────┘  ║  │
│  ║                                                                        ║  │
│  ║   Deployment: Self-Hosted (free) │ Cloud Sync ($3) │ Web Access ($20) ║  │
│  ╚═══════════════════════════════════════════════════════════════════════╝  │
│                                                                              │
└─────────────────────────────────────────────────────────────────────────────┘

MCP Server

"Bluetooth for AI" - universal memory access for any AI tool.

• memory_add
• memory_search
• memory_recall
• inbox_add

REST API v1

Full CRUD operations for entries, memories, and search.

• POST /entries
• POST /search
• POST /memories
• GET /review/:type

SDK Packages

Native libraries for popular languages.

• npm: @onelist/sdk
• pip: onelist-sdk
• hex: onelist

Integration	Effort	Impact	Priority	Notes
OpenClaw ClawdHub Skill	Low	High	P0	Natural fit, 100K+ users
MCP Server	Medium	High	P0	Universal AI tool support
REST API v1	Medium	High	P0	Foundation for all integrations
Twin Plugin	Medium	Medium	P1	Business automation users
Claude Code Extension	Low	Medium	P1	Developer audience
SDK Packages	Low	Medium	P1	Developer experience
ChatGPT Plugin	Medium	Medium	P2	Broad reach, complex approval
Supermemory Backend Option	Medium	Low	P3	Optional premium feature

Integration	Free Includes
OpenClaw Skill	Unlimited local memories
MCP Server	Self-hosted endpoint
Twin Plugin	Basic storage
REST API	1,000 calls/month

Ecosystem Integration Strategy

⚡ Executive Summary

+ Twin

+ Supermemory

+ OpenClaw

Platform Positioning Map

Capability Matrix

Twin Integration

What Twin Does Well

What Twin Users Need

Integration Architecture

Implementation: Twin Workflow Using Onelist

Supermemory Integration

What Supermemory Does Well

Where Onelist Complements

Market Segmentation

Supermemory Wins

Onelist Wins

MCP Interoperability

OpenClaw Integration

"OpenClaw does. Onelist remembers."

What OpenClaw Does Well

What OpenClaw Users Need

Integration Architecture

ClawdHub Skill Implementation

Quick Install

Example Usage

User Experience Example

Onelist as the Universal Memory Layer

MCP Server

REST API v1

SDK Packages

Implementation Priority

Success Metrics (6-Month Targets)

Revenue Model

Free Tier Value

Upgrade Triggers

Ready to Integrate?

Detailed Comparisons

Onelist vs Twin

Onelist vs Supermemory

Onelist + OpenClaw