Dispatch

Execution layer for autonomous AI agents

[Dispatch @dispatch_relay]

Dispatch is about to launch on Robinhood Chain.

The token is planned as the settlement token for agent execution and computation across Dispatch.

https://t.co/BLDayRwFoS

[Dispatch @dispatch_relay]

What happens between defining an AI agent and getting a useful result?

Dispatch is building a workflow around four layers: configuration, data, execution, and records. Here’s how they fit together and what you can build with custom agents 👇

[Dispatch @dispatch_relay]

In the wallet-audit agent (which functions like a forensic investigator), it combines on-chain data with verified threat. At the core, it connects directly to different nodes and RPCs to retrieve verifiable information, decoded transaction traces, spender allowances, and live simulations.

It answers risk-related questions like:
- Which dApps or contracts still have permission to move my funds?”
-Can you show me exactly what this transaction will do before I sign?”
-Is this yield farm actually paying out, or is it a slow rug?
- Has this contract been exploited before?

... and many more

Every claim links back to a traceable evidence ID, tested against a strict evaluation set of real exploits, revoked spenders, and edge-case wallets before touching mainnet.

[Dispatch @dispatch_relay]

Prior to mainnet and the launch of the platform token (will be launching today), we are rolling out two agents on Dispatch: quick-ask and wallet-audit.

https://t.co/Aok7kDPQDB

Here is what each agent does and how they function:

1. quick-ask (Rapid Testing & Prototyping) The built-in default agent designed for immediate, zero-setup testing in our console playground or via API. It is a lightweight single-turn inference with fast response times. Used to test prompts, benchmark hardware latency across GPU nodes, and verify our execution receipts without needing to configure tools or pre-fund a custom agent.

2. wallet-audit (Autonomous On-Chain Security) A specialized risk analysis agent built for deep wallet intelligence and security audits. It connects directly to blockchain RPCs for. calculations (P&L, allowance exposure, portfolio metrics) are handled deterministically in code

[Dispatch @dispatch_relay]

Prior to mainnet and the launch of the platform token (will be launching today), we are rolling out two agents on Dispatch: quick-ask and wallet-audit.

https://t.co/Aok7kDPQDB

Here is what each agent does and how they function:

1. quick-ask (Rapid Testing & Prototyping) The built-in default agent designed for immediate, zero-setup testing in our console playground or via API. It is a lightweight single-turn inference with fast response times. Used to test prompts, benchmark hardware latency across GPU nodes, and verify our execution receipts without needing to configure tools or pre-fund a custom agent.

2. wallet-audit (Autonomous On-Chain Security) A specialized risk analysis agent built for deep wallet intelligence and security audits. It connects directly to blockchain RPCs for. calculations (P&L, allowance exposure, portfolio metrics) are handled deterministically in code

[Dispatch @dispatch_relay]

We also want to extend custom agents with controlled tool integrations.

Examples:
• Read an approved data source
• Query blockchain state
• Retrieve relevant documents
Each integration needs clear permissions and traceable results. Custom prompts alone don’t grant these capabilities.

[Dispatch @dispatch_relay]

What happens between defining an AI agent and getting a useful result?

Dispatch is building a workflow around four layers: configuration, data, execution, and records. Here’s how they fit together and what you can build with custom agents 👇

[Dispatch @dispatch_relay]

Every completed run has a record.
Inspect its input, output, duration, token usage, routing source, and saved receipt.
Wallet investigations also store an evidence digest. It identifies the saved snapshot; it isn’t an on-chain attestation or a guarantee of safety.

[Dispatch @dispatch_relay]

Dispatch allows users to use define agents, configure reusable models, system instructions, and tool sets.

Workloads are dispatched directly to cloud GPU nodes across the network. Every completed run returns an immutable receipt with exact hardware provenance, latency, and token metrics. For settlement, workspaces are funded through prepaid smart contract vaults powered by our token (token will be launched today).

Read more at: https://t.co/DG4j8b7w6k

Live on Robinhoodtestnet

Testnet Settlement token (DSM)
https://t.co/4GbDoK8BM5

Credit vault (deposits → prepaid credits)
https://t.co/KiiaALuAzQ

Protocol treasury
https://t.co/FIceOGQSVM

[Dispatch @dispatch_relay]

Dispatch

[Dispatch @dispatch_relay]

Meet Dispatch

Define AI agents, run your tasks, and track results and costs, all in one workspace.
We’re building toward a simpler way to submit work, without managing model providers, GPU infrastructure, billing, and execution records separately.