Vilix

Works with

• Claude
• Codex
• ChatGPT
• Cursor
• OpenClaw
• Hermes
• Gemini Spark
• Claude Code
• Grok
• Grok Build
• Manus
• GitHub Copilot
• Windsurf
• Devin
• Lovable
• Muse
• Cline
• Continue
• Gemini CLI

Let your other AI tools explain the memory layer

Each link opens a new chat with a prompt anchored to vilix.ai, so search-capable models read the site instead of guessing from stale memory.

Social proof

Builders are talking about Vilix AI

Real feedback from people using Vilix AI every day.

Mati Nemera Software Engineer, Spectrum

From the founder

I used Vilix to build Vilix. In 10 days.

I'm Abdeljabbar, the founder. Code, marketing site, pricing, design, all shipped in 10 days. I dogfooded the product the entire way.

Every AI tool I used, OpenClaw, Hermes, Claude Code, Cursor, Codex, ChatGPT, was plugged into Vilix. Same memory across all of them. Laptop and phone.

Plan a feature on my phone at a coffee shop. Open my laptop. Cursor already knows. No re-explaining. No copy-paste. Just continue where I left off, every time.

If it worked to ship Vilix, it'll work for whatever you're building.

Abdeljabbar Elassali · Founder, Vilix (Appfairly LLC)## Part of Microsoft for Startups

The program provides Vilix AI with Azure and AI credits without taking equity.

More than just chat history

Vilix doesn't only remember your conversations. It saves your projects, tasks, the rules every AI should follow, and the skills that teach agents how you do things, so every connected tool sees the full picture of how you work.

user_rules are included on every plan, with up to 20 personal rules per account. Project boards include project_rules per board: stack, conventions, and tone supplied to connected clients. And Agent Skills turn recurring procedures into steps your connected agents can load and follow. With the Agent Inbox you can push work to your own running agents from any client, and they pick it up on their next memory read. See all features.

Coming soon

Tables: memory your agents manage

We're building the next Vilix AI upgrade: decision logs, scheduled run histories, and tasks as structured tables your agents create and maintain. Join the waitlist to shape what ships first and get early access.

What is Vilix AI?

One shared memory for all your AI tools.

1

Connect Vilix AI

Add Vilix AI to each client using OAuth or an API key, then enable the auto-memory instructions.

2

Talk normally

Use Claude Code, Cursor, Codex, or ChatGPT like usual.

3

Continue anywhere

Vilix retrieves relevant context when you switch tools.

Works with the AI you already use.

Works across your AI tools

• OpenClaw
• Hermes
• Claude Code
• Cursor
• Codex
• ChatGPT
• Claude
• Grok
• Manus AI
• GitHub Copilot
• Windsurf
• Lovable
• Muse
• Any MCP-compatible AI

Works everywhere you use AI

• Desktop
• Mobile (same account, same memory)
• Browser & cloud tools

Connect each client to one account, then use saved context across your tools.

Connect a headless agent

Running a headless agent like Hermes or OpenClaw on a box, with no browser for OAuth? It connects with two things:

1. 1.The Vilix MCP URL, https://api.vilix.ai/mcp.
2. 2.An API key sent as a Bearer header. No browser OAuth.

Generate the key on the Agents page in your Vilix dashboard.

Why it matters

Why Vilix is different

Native memory helps inside its own product. Vilix AI gives separately connected clients a shared store for saved conversations and agent work state.

	Vilix AI	ChatGPT Memory	Claude Memory & Projects
Memory scope	One Vilix account across connected clients	ChatGPT account and supported connected context	Claude memory plus separate project context
Past conversations	Retrieves conversations saved to Vilix	Can use past chats and other available context	Memory and past-chat search; project boundaries apply
Source attribution	Saved source and conversation metadata	Memory sources can show personalization context	Past-chat citations link to original conversations
Setup	Connect each client; enable memory instructions	Built-in settings; availability varies	Built-in memory and project settings

Direct competitor snapshot

If you are deciding between Vilix AI and direct alternatives, compare these current product surfaces: the Mem0 ecosystem and Supermemory MCP.

Capability	Mem0 ecosystem	Supermemory MCP
MCP-native connect	Check the specific Mem0 integration; OpenMemory is now a session-porting CLI	Hosted MCP with OAuth
Automatic capture	API calls or product-specific capture; extension repository is archived	Assistant calls save/forget tools; ingestion and uploads also available
Structured projects & tasks	Platform projects organize API work; not equivalent to task boards	Spaces organize documents and memories; not equivalent to task boards
Works on mobile	Depends on the product and client	Depends on the MCP client

Want the full breakdown? Compare all direct alternatives or read the full matrix on cross-AI memory.

Data control

You own your memory

• Export all your data anytime.
• Delete your entire account instantly.
• No selling user data.
• No training on your private data.
• OAuth or API-key access, scoped to your account.

Frequently asked questions

Does this work on mobile?

Mobile support depends on the client. Claude supports connected services on mobile, with connector installation in beta; its web or desktop setup is the fallback. Gemini Spark can be connected in a phone browser and used in its mobile app. Terminal agents run on their configured computer or server. Check the device-specific guide before connecting.

Which AI tools are supported?

Claude, Claude Code, Codex, Cursor, OpenClaw, Hermes and other clients listed in the setup guide. Each client needs compatible transport, authentication and tool permissions; MCP support alone does not verify every integration.

Do I need a paid Claude or ChatGPT plan?

Claude custom connectors are available on Free, Pro, Max, Team and Enterprise; Free allows one custom connector. ChatGPT Developer mode requires Plus, Pro, Business, Enterprise or Edu on the web. Other clients have their own account, model and administrator requirements, separate from your Vilix AI plan.

Anything special for ChatGPT?

On an eligible ChatGPT web account, enable Developer mode under Settings → Security and login. Go to Plugins, select +, add Vilix AI with OAuth, then enable it in the conversation. Workspace permissions and tool approvals still apply.

Does Vilix work with Gemini?

Gemini CLI and Gemini Spark have separate setup guides. This is not blanket support for consumer Gemini chat. Spark custom apps require eligible Google accounts and access to Spark; memory saves may need confirmation. Gemini CLI eligibility changed in June 2026, so check its current access requirements.

What are Agent Skills?

Reusable how-to procedures saved once in your Vilix memory. Write down how a recurring task is done, like a "check active users" skill with the exact query and what counts as active, and every connected AI loads the procedure on demand and runs it the right way. Scope a skill globally or to a single project, and edit it from any tool.

Do I need to configure each tool?

Yes. Connect each AI tool to the same Vilix AI account, enable its memory tools, and add the auto-memory instructions. One approval does not configure your other clients. Once connected, they can retrieve the context saved to that shared account.

Does Vilix AI save conversations or only extracted facts?

Vilix AI stores the user and assistant exchanges your clients send through save_turn, with source and conversation metadata. It also derives useful memories from those exchanges. Retrieval selects relevant saved context; it does not send your entire archive into every prompt. Connecting a client does not automatically import its old chat history, and attachment summaries are not a copy of the original files.

Can I export my data?

Yes, anytime. You can pull all of your memory out of Vilix in a portable format whenever you want.

Can I delete everything?

Yes. You can delete individual memories or wipe your entire account instantly, with no waiting period.

Is my data private?

Your data is isolated per user. We do not sell user data and we do not train third-party models on your private memory. Access uses OAuth or a personal API key. Connected clients receive the context they retrieve.