Making AI agents actually useful
ContextQ is built by Henia — a small team making AI agents more useful in real-world workflows. The problem is simple: AI agents are powerful but stateless. Every conversation starts from zero. ContextQ gives them a shared, persistent memory layer.
The problem we're solving
Developers using Claude Code, Cursor, or Copilot have to re-explain context every session — project decisions, naming conventions, team patterns, past debugging findings. This context lives in human heads, not in the tools.
You tell your AI agent that your team uses camelCase, that deployments go through EasyPanel, that the auth middleware was rewritten for compliance. Then the session ends. Next time, you start over. Every time.
We're fixing that. ContextQ stores what your agents learn so they never forget. The knowledge compounds across sessions, across projects, and across your entire team.
How it works
ContextQ is an MCP server that any AI agent can connect to. It stores context entries — references, feedback, lessons learned, project decisions — in a structured database with hybrid search powered by full-text indexing and vector similarity.
When your agent starts a new session, it searches ContextQ for relevant context. When it learns something new, it saves it back. Knowledge flows between sessions automatically.
Save
Agents store what they learn
Search
Hybrid full-text + vector retrieval
Dream
AI consolidates and surfaces insights
Built by Henia
Henia is a small software studio based in Vietnam. We build tools for developers and AI-powered workflows. ContextQ is our first public product.
We started ContextQ because we hit the same problem ourselves — re-explaining context to AI agents across every project, every session. Once we built the solution internally, we realized every team using AI coding tools needs this.
Small team, focused product. No venture funding, no bloated feature list. Just the thing that makes your agents remember.
Visit heniart.com9
MCP tools
3
Search engines
2026
Launched April
Open source and roadmap
The core MCP protocol integration is open. We believe the protocol layer should be transparent so developers can inspect, extend, and trust what their agents are doing.
What's next
- Payment integration — Stripe billing for Pro and Enterprise plans
- More agent clients — First-party integrations for Cursor, Windsurf, and Cline
- Self-hosted Docker image — Full ContextQ stack for Enterprise on-premise deployment
- Team dashboards — Usage analytics, context health monitoring, and entry management UI
Ready to give your agents memory?
Start for free. No credit card required.