Oros brings AI-driven interactions to any dApp. Designed as an intuitive agent layer, Oros transforms how users interact with blockchain ecosystems by bridging advanced AI with on-chain capabilities. Imagine managing your assets, executing complex DeFi strategies, or exploring new dApps—all with a simple conversation. That’s the vision of Oros.
Overcoming technical complexity and fragmented user experiences is a key unlock for web3 adoption. Oros simplifies blockchain interaction, introducing a consistent, approachable AI persona—your trusted guide across dApps.
- Natural Language → On-Chain Actions: Oros translates user queries and prompts into blockchain function calls, letting your users “speak” directly to DeFi, NFTs, governance, or any on-chain logic.
- Smart Context: Oros provides a unified, intelligent presence that feels personalized and seamless across web3, whether you’re staking tokens or exploring yield strategies.
- Universal Access: By integrating with decentralized applications, Oros democratizes access to blockchain intelligence, making Web3 easier for everyone.
For Users
Blockchain interactions often feel complex. With Oros, you can simply say:
- “Stake my KAVA rewards”
- “Send 100 USDT to Alice”
Oros understands your intent, navigates the technicalities, and executes securely.
For Developers
Integrating Oros means less time spent building custom AI tools and more time focusing on what your dApp does best. A straightforward SDK enables any dApp to embed the Oros experience effortlessly.
+------------------------------+
| Oros UI (Front-End) |
| (React/Typescript) |
+---------------+--------------+
|
▼
+---------------------------------------+
| Oros Agent (Chat + Tools Orchestr.) |
| - Core AI logic (LLM-based) |
| - Tool registry (function specs) |
| - Session & memory layer |
+-------------------+-------------------+
|
▼
+---------------------------------------+
| Tools / Skills |
| - Contract calls, off-chain APIs |
| e.g., stakeTokens(), transferERC20() |
+-------------------+-------------------+
|
▼
+--------------------+
| Blockchain(s) |
| (Kava, EVM, etc.) |
+--------------------+
- Front-End: The user-facing interface (React) that embeds an Oros chat widget.
- Oros Agent: Receives user prompts, uses LLM logic & memory, decides which tool(s) to call.
- Tool Registry: A set of typed “function calls.” Each function can be a contract method or off-chain API.
- Blockchain: Actual transaction logic — Kava chain, EVM, or custom protocols.
At its core, Oros is a bridge between natural language and blockchain actions, leveraging cutting-edge AI to interpret user requests and perform on-chain operations.
-
Understanding Intent
Oros processes user input using advanced language models, breaking down requests into actionable components. Whether it’s staking tokens or bridging assets, Oros knows what needs to be done. -
Executing Actions
Oros translates user intents into on-chain function calls, working within the frameworks of supported dApps. Every action is verified and secure. -
Building Relationships (Next Steps)
Over time, Oros learns user preferences and habits (securely and with consent), providing a richer, more tailored experience.
-
Conversational Simplicity
Interact with blockchain assets as if you’re chatting with a knowledgeable friend. -
Universal Compatibility
Works across multiple dApps and ecosystems, starting with the Kava blockchain. -
User-Centric Security
Every transaction requires user approval, ensuring complete control over actions. -
Developer-Friendly Integration
An easy-to-use SDK makes embedding Oros into any dApp a breeze.
Milestone | Description | Status |
---|---|---|
Oros MVP | Basic chat + tool bridging, single dApp integration | ✅ Done |
Multi-dApp Integration | Expand Oros to multiple Kava dApps | 🚧 In progress |
Cross-Chain Support | EVM bridging (ERC20 transfers, bridging logic) | 🚧 In progress |
Advanced Memory | Persistent user context across sessions | 📅 Planned |
TEE / Security | Integrate secure enclaves for private data + signing | 📅 Planned |
deModel Integration | Connection to community-trained LLMs | 📅 Planned |
Before you begin, ensure you have the following tools installed:
- Go: Install Go (version 1.19 or higher recommended)
- Node.js and npm: Install Node.js (version 18.x or higher recommended)
To verify installations, run:
go version
node -v
npm -v
Clone the Oros project to your local machine:
git clone https://github.com/Kava-Labs/oros.git
cd oros
Oros requires an OpenAI API key for its backend. Obtain your key from OpenAI's API portal.
Run the following command to start the backend API:
OPENAI_API_KEY=<your_key> OPENAI_BASE_URL=https://api.openai.com/v1 go run ./api/cmd/api/main.go
Before starting the frontend, create a .env
file in the root of the project to configure the necessary environment variable. The file should contain the following:
VITE_OPENAI_BASE_URL=http://localhost:5555/openai/v1
After creating the .env
file, install the required dependencies by running:
npm install
npm run dev
The frontend should now be running on http://localhost:3000
. Open your browser and interact with Oros locally!
If you encounter issues:
- Ensure your OpenAI API key is valid.
- Verify all required tools (Go, Node.js, npm) are installed and up-to-date.
- Check for detailed error logs in the backend or frontend terminal output.