For Developers
- Simple REST API and official SDKs (TypeScript, Python)
- JSON metadata, tags, groups for flexible querying
- Works with any stack that can do HTTP
For AI Agents
- Perfect for Claude Code and Copilot via MCP
- Store conversation history, decisions, and user preferences
- Hybrid semantic + keyword search tuned for agent memory, not generic search
For Privacy-Sensitive Teams
- Embeddings generated locally with Transformers.js; nothing sent to OpenAI/Anthropic
- Data stored in standard PostgreSQL + pgvector; no proprietary vector DB lock-in
- No external dependencies on AI providers for core functionality
See PersistQ in Action
45-second demo
One prompt to setup — watch how PersistQ stores, searches, and retrieves memories in real-time
Built for Developers, Optimized for AI
Everything you need to add semantic memory to your AI agents.
No external API calls for embeddings
Generate embeddings locally with Transformers.js (384d). Your data never leaves your infrastructure.
Find the right memory instantly with lightning-fast semantic search
Powered by pgvector with cosine similarity. Query your memories naturally and get relevant results every time.
Combine semantic and exact matching for the most relevant results
Hybrid keyword + vector search ensures you never miss important memories, whether the query is fuzzy or exact.
Fast enough for real-time agent loops
Average latency of ~200ms means your AI agents never wait. Optimized for production workloads at scale.
Works with AI tools.
MCP ready.
Integrates seamlessly with Claude Code and GitHub Copilot CLI via Model Context Protocol. Simple setup, powerful memory.
- Claude Code integrationFull MCP support with tools and resources
- GitHub Copilot CLIMCP tools integration available
- Quick setupInstall globally and configure in minutes
Three ways to integrate
Choose the integration that works best for your workflow
12345678910111213141516171819import { createClient } from 'persistq-sdk'; const client = createClient({ baseUrl: 'https://api.persistq.com', apiKey: process.env.PERSISTQ_API_KEY,}); // Store a memoryawait client.createMemory( 'User prefers dark mode', 'preferences', { tags: ['ui', 'settings'] }); // Search memoriesconst results = await client.searchMemories( 'user preferences', { limit: 10 });Use Cases
Add persistent memory to any AI application
AI Assistants
Give your assistant long-term memory of preferences, tasks, and interactions.
Example: Claude remembers your tech stack (Next.js + PostgreSQL + Tailwind) and stops suggesting React class components or MongoDB.
Customer Support Agents
Add a persistent knowledge layer to multi-step agents.
Example: Your AI support agent remembers each user's last 5 issues and doesn't ask the same onboarding questions again.
Custom AI Apps
Store contextual data, notes, summaries, and conversation history.
RAG Enhancements
Improve retrieval quality with structured semantic memory.
From the Founder
PersistQ is new — but it was built from real frustration with privacy concerns, overly complex vector databases, and the need to send data to third parties for embeddings.
Developers deserve a memory layer that is fast, private, simple, and easy to integrate.
PersistQ gives you semantic memory with local embeddings, no vendor lock-in, and full Claude MCP support. No external API calls for embeddings means your data never leaves your infrastructure.
If you're building an AI agent, PersistQ is designed for you.
Simple, transparent pricing
Start free. Scale as you grow. No hidden fees.
Powerful Features
Everything you need to build AI applications with persistent memory
Average search response
Memories per project
Uptime guarantee
No external dependencies
JSON metadata support
Understands meaning and context
Create, read, update, delete
Organize by topic/use case
Data protection at rest
Embeddings never leave your system
Privacy by design
Your data, your rules
Full type support
Standard HTTP endpoints
Guides and examples
Works with any tech stack