🦞 Give OpenClaw superpowered memory with one command
Install Skill
All Comparisons

Onelist

Privacy-first memory platform

vs
🧠

Supermemory

Universal Memory API

Two different approaches to AI memory. Supermemory is a memory API for developers with state-of-the-art benchmarks. Onelist is a complete privacy-first platform with agents and UI.

TL;DR

Choose Onelist if you want:

  • ✓ A complete PKM solution with Web, Desktop, Chat, and API
  • ✓ Atomic extraction + two-layer retrieval (Reader & Searcher Agents)
  • ✓ Asset enrichment (transcription, OCR, action items from meetings)
  • ✓ Content capture (Web Clipper, imports, cloud storage sync)
  • ✓ Self-hosted with E2E encryption, or cloud options
  • ✓ Full GTD life management (River Agent)
  • ✓ Task automation via webhooks, MCP tools, internal tools
  • ✓ Open source (MIT) with predictable pricing

Choose Supermemory if you want:

  • ✓ State-of-the-art benchmark performance (81.6% LongMemEval)
  • ✓ Sub-300ms retrieval latency at scale
  • ✓ Memory API for your own AI application
  • ✓ Enterprise-proven infrastructure (5B tokens/day)
  • ✓ MCP server for portable memory across AI tools
  • ✓ Intelligent forgetting and decay curves

Memory Performance Benchmarks

LongMemEval Benchmark Results

LongMemEval tests real-world memory capabilities: multi-session reasoning, temporal understanding, and knowledge updates.

81.6%
Supermemory Overall
71.4%
Multi-Session
76.7%
Temporal Reasoning
~55%
Industry Average

Onelist Status: Reader Agent (atomic extraction with reference resolution) and Searcher Agent (two-layer retrieval with similarity check) are implemented. We're targeting comparable benchmark performance. Results will be published when LongMemEval testing is complete.

Supermemory Performance

  • Retrieval Latency: Sub-300ms
  • Scale: 50M tokens per user
  • Throughput: 5B tokens/day (production)
  • vs. Competitors: 10x faster than Zep, 25x faster than Mem0

Onelist Performance

  • Retrieval Latency: Sub-500ms target
  • Scale: Self-hosted dependent (PostgreSQL)
  • Architecture: PostgreSQL + pgvector + Searcher Agent
  • Advantages: No cloud dependency, full control, no token costs
  • Trade-off: You manage infrastructure, but zero egress fees

Feature Comparison

Capability Onelist Supermemory Winner
Platform Type Complete PKM solution Memory API only Onelist
User Interface Web, Desktop, Chat (River), REST API None (API only) Onelist
Self-Hosting Yes (free forever) No (cloud only) Onelist
E2E Encryption Yes (Cloud Sync tier) No Onelist
Atomic Memory Extraction Yes (Reader Agent) Yes (core feature) Both
Two-Layer Retrieval Yes (Searcher Agent) Yes (core feature) Both
Asset Enrichment Transcription, OCR, descriptions, action items None Onelist
Content Capture Web Clipper, imports (10+ sources), cloud storage API ingestion, connectors Onelist
Benchmark Performance TBD (targeting SOTA) 81.6% LongMemEval Supermemory
Retrieval Latency Sub-500ms target Sub-300ms Supermemory
Intelligent Forgetting Planned Built-in decay curves Supermemory
MCP Server Planned (post-MVP) Yes (portable memory) Supermemory
Task Automation Internal tools, webhooks, MCP tools None Onelist
Agent Ecosystem River, Reader, Searcher, Feeder, Asset Enrichment None Onelist
Life Management Full GTD (buckets, contexts, reviews) None Onelist
Open Source Full platform (MIT) Partial (benchmarks only) Onelist

Key Differences Explained

Platform vs. API

Fundamental Difference

Onelist: Complete Platform

Onelist is a full-stack personal knowledge management solution:

  • • Web app for browser access
  • • Desktop app for native experience
  • • Chat interface via River Agent
  • • REST API for programmatic access
  • • 5+ specialized agents (Reader, Searcher, Feeder, Asset Enrichment, River)

Best for: End users who want a complete solution with UI, OpenClaw users, people consolidating from Evernote/Notion/Obsidian.

Supermemory: Infrastructure API

Supermemory is memory infrastructure - an API that developers use to add memory capabilities to their own AI applications. No user interface included.

Best for: Developers building AI apps, teams who want to integrate memory into custom products, enterprises with existing applications.

Memory Architecture

Both Strong

Onelist's Implementation

Reader and Searcher Agents implement Supermemory-inspired patterns:

  • Reader Agent: Atomic memory extraction with reference resolution ("She" → "Sarah mentioned...")
  • Searcher Agent: Two-layer retrieval - search atomic memories for precision, return source chunks for context
  • Similarity Check: Pre-storage duplicate detection prevents redundant memories
  • Temporal Metadata: content_created_at + ingestion time for temporal queries
  • Knowledge Graph: Entries link to people, projects, topics with relationship tracking

Roadmap: Intelligent forgetting, decay curves

Supermemory's Innovations

  • Atomic Memory Extraction: Breaks content into discrete, searchable facts with resolved references.
  • Two-Layer Retrieval: Search on memories (high precision), return with source chunks (context).
  • Bi-Temporal Model: Tracks both event time and ingestion time for temporal reasoning.
  • Intelligent Forgetting: Decay curves reduce weight of old, unused memories.
  • Graph Relationships: Supersedes/Refines/Contradicts relationship tracking.
  • Benchmarked: 81.6% LongMemEval - proven at scale

Asset Enrichment & Content Capture

Onelist Advantage

Onelist: Full Media Intelligence

Asset Enrichment Agent transforms raw uploads into searchable knowledge:

  • • Audio: Whisper transcription with timestamps, speaker diarization, action item extraction
  • • Images: AI descriptions, OCR, face recognition linked to person entries
  • • Video: Keyframe analysis, audio extraction, chapter generation
  • • Documents: PDF text extraction, table parsing

Feeder Agent captures content from: Web Clipper, Evernote, Notion, Obsidian, RSS, cloud storage (Google Drive, Dropbox, OneDrive), and more.

Supermemory: API-Only

Supermemory is a memory API - you send text, it stores and retrieves memories. No asset processing built in.

Content capture via connectors (Google Drive, Notion, OneDrive, S3) but no enrichment - audio, images, and documents remain as-is.

If you need transcription or OCR, you must integrate separate services before sending to Supermemory.

Privacy & Data Control

Onelist Advantage

Onelist: Full Control

  • Self-Hosted: Run on your own infrastructure, free forever
  • E2E Encryption: Cloud sync encrypted so we can't read your data
  • BYOB Storage: Use your own S3-compatible bucket
  • Open Source: Audit the code, no vendor lock-in

Supermemory: Cloud-Only

  • No Self-Hosting: All data processed on Supermemory servers
  • Trust Required: Must trust their data handling
  • SOC 2 Compliant: Enterprise security certifications
  • Partial Open Source: Benchmarks open, core proprietary

Pricing Comparison

Onelist Pricing

Self-Hosted
$0 forever
Cloud Sync
$3/mo + usage
Web Access
$20/mo

Predictable pricing, no token surprises.

Supermemory Pricing

Free Tier
1M tokens/mo
10K queries/month
Overage
$0.01/1K tokens
$0.10/1K queries
Enterprise
Custom

Great for experimentation, can get expensive at scale.

Cost Example: Heavy User (100M tokens/month)

$20/mo
Onelist Web Access
~$900/mo
Supermemory (estimated)

Supermemory-Inspired Architecture

Already Implemented (MVP)

  • ✓
    Atomic Memory Extraction - Reader Agent breaks content into discrete, searchable facts with resolved references
  • ✓
    Two-Layer Retrieval - Searcher Agent searches atomic memories for precision, returns source chunks for context
  • ✓
    Similarity Check - Pre-storage duplicate detection prevents redundant memories
  • ✓
    Temporal Metadata - content_created_at + ingestion time for temporal queries
  • ✓
    Knowledge Graph - Entries link to people, projects, topics with relationship tracking

On Roadmap (Post-MVP)

  • →
    MCP Server - Expose Searcher Agent as MCP server for portable memory across AI tools
  • →
    Intelligent Forgetting - Decay curves to reduce weight of old, unused memories
  • →
    LongMemEval Benchmarking - Publish benchmark scores to validate performance

Why Onelist Is the Right Choice

Supermemory-class memory architecture plus capabilities they don't offer:

  • ✓ Asset Enrichment - Transcription, OCR, descriptions, action items from meetings
  • ✓ Content Capture - Web Clipper, imports from 10+ sources, cloud storage sync
  • ✓ Full UI Stack - Web, Desktop, Chat interface - not just an API
  • ✓ GTD Life Management - River Agent with buckets, contexts, weekly reviews
  • ✓ Task Automation - Internal tools, webhooks, MCP tool execution
  • ✓ Self-Hosted - Full control, E2E encryption, no cloud dependency
  • ✓ Open Source (MIT) - Audit the code, no vendor lock-in
  • ✓ Predictable Pricing - No token surprises at scale

Ready to Try Onelist?

Get Supermemory-class memory capabilities with privacy, self-hosting, and a complete application stack.

Get Started Free OpenClaw Integration