5 parts of Casper’s AI toolkit for builders
5 parts of Casper’s AI toolkit show how agents can pay, build, and deploy on mainnet today.

Casper’s new toolkit lets AI agents pay for services and ship contracts on mainnet.
Casper Network says its AI Toolkit is live on mainnet now, with a production x402 Facilitator and a $150,000 buildathon backing it.
| Item | What it does | Live status |
|---|---|---|
| x402 Facilitator | Settles HTTP-based micropayments for paid requests | Live on mainnet |
| MCP support | Lets agents query chain data and invoke tools | Live on mainnet |
| CSPR.trade MCP Server | Lets agents swap, provide liquidity, and manage portfolios | Live on mainnet |
| Odra Framework | Helps agents write, test, and deploy smart contracts | Available in toolkit |
| CSPR.build Agent Skills | Adds wallet, signing, event, and API integration | Available in toolkit |
1. x402 Facilitator
Get the latest AI news in your inbox
Weekly picks of model releases, tools, and deep dives — no spam, unsubscribe anytime.
No spam. Unsubscribe at any time.
The x402 Facilitator is the payment rail that makes Casper’s agent story concrete. An agent can call a paid endpoint, receive an HTTP 402 response with a price, sign authorization, and get the data back in seconds. No API keys or human billing flow are required.

Casper says this is the first WebAssembly-native Layer 1 with live HTTP-based micropayment infrastructure for AI agents. That matters because the payment happens inside the request itself, which fits machine-to-machine workflows better than subscription-style billing.
- Protocol: x402
- Payment style: pay per request
- Settlement: on-chain
- Status: live on mainnet
2. Model Context Protocol support
Through Model Context Protocol, or MCP, AI agents can discover and use external tools on Casper. That includes querying balances, submitting transactions, reading contract state, and watching transfer events as they happen.
This is the part of the stack that turns an agent from a text generator into an on-chain operator. It gives the model a standard way to reach into the network and act on current data instead of stale snapshots.
- Balance checks
- Transaction submission
- Contract-state reads
- Real-time event streams
3. CSPR.trade MCP Server
The CSPR.trade MCP Server expands the agent’s role into DeFi. With natural language, an agent can swap assets, provide liquidity, and manage a portfolio on Casper without switching tools or waiting for a human to click through a dashboard.

For teams building trading bots, treasury agents, or portfolio assistants, this is the most direct example in the release of an AI system acting like a market participant. It is not just reading chain data, it is using that data to move funds and position assets.
- Swaps
- Liquidity provision
- Portfolio management
- Natural-language control
4. Odra Framework
Odra is Casper’s smart contract toolkit for the build side of the stack. The company says an AI coding agent can point at Odra and go from idea to a deployed contract without human guidance, including writing and testing along the way.
That makes Odra the bridge between agentic productivity and production software. If your goal is to have software create software, this is the component that closes the loop from prompt to deployment.
AI agent -> Odra -> write contract
AI agent -> Odra -> test contract
AI agent -> Odra -> deploy contract
5. CSPR.build Agent Skills
CSPR.build Agent Skills package the rest of the integration work into one installable bundle. Casper says it includes wallet connections, transaction signing, event handling, and CSPR.cloud API access, backed by REST, Streaming, and Node API layers.
For builders, this is the glue that makes the toolkit usable at scale. It reduces the amount of custom wiring needed when an agent has to read from the network, write to the network, and keep track of what changed.
- Wallet connections
- Transaction signing
- Event handling
- CSPR.cloud API access
How to decide
If you are building payment-aware agents, start with x402 Facilitator first. If your agent needs to inspect chain state or call tools, MCP support is the right entry point. If you want the agent to trade, CSPR.trade is the most specific fit.
If your goal is autonomous app creation, Odra and CSPR.build belong together. Odra handles the contract workflow, while CSPR.build fills in the wallet, signing, and API pieces that turn a prototype into something that can run on Casper mainnet.
// Related Articles
- [IND]
Zhihu users map out 2024 with family, work, travel
- [IND]
5 ways Kimi Work helps knowledge workers
- [IND]
5 reasons to use Endive on the JVM
- [IND]
Why LibreOffice’s browser push is the right move
- [IND]
AI Slop Is Flooding Music Streaming Apps
- [IND]
Why RISC-V and GPU Pairing Is the Right SoC Bet