Give your AI agent persistent, searchable memory that you can see, edit, and trust.
Every conversation starts from scratch. Context is lost.
Good luck finding that one thing from three weeks ago.
It's a black box. You have no control or visibility.
Your notes are separate from your AI's context.
Both you and your AI read and write to the same knowledge base.
Agents store memories programmatically through a clean REST API.
You have full visibility. Edit, tag, and organize AI memories.
Hybrid search works for humans and AI alike.
┌─────────────┐ ┌─────────────┐ ┌─────────────┐
│ │ │ │ │ │
│ You │◄───────────────►│ Onelist │◄───────────────►│ AI │
│ │ Read/Write │ │ Read/Write │ Agent │
│ │ │ │ │ │
└─────────────┘ └─────────────┘ └─────────────┘
│
│
┌───────────┴───────────┐
│ │
│ Same entries. │
│ Same search. │
│ Shared understanding│
│ │
└───────────────────────┘
Not all memories are equal. Onelist organizes memory into four layers for optimal retrieval and cost efficiency.
Core facts that never change. Always loaded.
Name, timezone, critical rules
Dynamic preferences. Topic-based loading.
Communication style, patterns
Recent context. Recency + relevance.
Recent conversations, threads
Generated on-demand. Temporary.
Synthesized answers, insights
Total context budget: ~1,200-3,800 tokens
vs. unbounded loading in naive approaches
Multiple representations of the same content let you optimize for cost without sacrificing quality.
Load full content for every memory
100 memories x 500 tokens = 50,000 tokens
~$0.50 per request
Load summaries, fetch full only when needed
100 memories x 50 tokens = 5,000 tokens
~$0.05 per request
Moltbot remembers your preferences, past conversations, and accumulated knowledge across sessions.
AI accumulates research findings that you can review, annotate, and build upon.
Your AI learns your preferences, writing style, and domain knowledge over time.
You and your AI collaborate on the same project documentation.
curl -X POST https://api.onelist.my/v1/entries \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{
"entry_type": "memory",
"title": "User prefers dark mode",
"content": "User mentioned they always use dark mode...",
"tags": ["preferences", "ui"],
"metadata": {
"source": "moltbot",
"confidence": 0.95
}
}'
Full API documentation available at /docs/api
Start building AI-powered applications with persistent memory today.