Trunk® — Accelerating MCPs through Smart Batching and Unified Delivery

Trunk is a Solana-native protocol architected for MCP-centric agents, leveraging Solana’s parallel runtime, intelligent request batching, and immutable decentralized storage to optimize agent performance at scale.

While we can’t predict every step toward AGI, we believe MCPs are a foundational leap, expanding agent capabilities, enabling true autonomy, and bringing us closer to building systems that can reason and act at scale.

Consistent performance for AI workloads, big and small.

MCP Call Pricing

Call Type

Cost

Search Request

0.001 $TRUNK

Schedule Meeting

0.001 $TRUNK

Predictable Costs

Fees priced per MCP call, not raw data size.

Request ID

#5781

Status

Confirmed

Confirmation Time

231 ms

Target

<400 ms

Blazing Speed

Confirmations under 400ms, even at peak load

MCP Traffic Overview

Requests

MCP Streams

128

Network Status

Optimal

Peak Capacity: 50,000+ MCP/sec

Massive Throughput

Tens of thousands of requests per second without saturating the network

Trunk delivers infrastructure with strict service-level guarantees, predictable performance, and machine-speed responsiveness — without sacrificing decentralization or cost-efficiency.

Sub-400ms Finality

Fast, reliable confirmation with latency guarantees built into the protocol.

Economical Microtransactions

Batch-optimized fees make high-frequency, low-value calls viable.

Massive Throughput

Handles tens of thousands of MCP calls per second without slowdown.

Trunk™ makes your agents more capable.

Read documentation

*Trunk Beta public release is coming Q3 2025

Follow Trunk and get updates.

Securely enable your employees and automate workflows with industry leading AI products. Batches thousands of MCP requests in real-time, optimizing for cost, latency, and batch variant

Secure by Design

End-to-end encrypted data handling ensures your workflows remain private.

Real-Time Batching

Intelligently grouped MCP calls to reduce compute costs and improve throughput.

Latency-Optimized

Engineered for machine-speed responsiveness, giving agents sub-second access to external tools and services.

Cost-Efficient at Scale

Smart batching and Solana-native architecture drive down operational costs without sacrificing reliability.

Secure by Design

End-to-end encrypted data handling ensures your workflows remain private.

Real-Time Batching

Intelligently grouped MCP calls to reduce compute costs and improve throughput.

Latency-Optimized

Engineered for machine-speed responsiveness, giving agents sub-second access to external tools and services.

Cost-Efficient at Scale

Smart batching and Solana-native architecture drive down operational costs without sacrificing reliability.

A Sustainable, Deflationary Economic Model. The TRUNK token powers the network’s staking, security, and governance layers

Sequencer

Rewards

Fee Burn

Supply & Inflation

ETH Restaking

Sequencer Rewards

70% of sequencer revenue goes to TRUNK stakers. The remaining 30% funds the ecosystem treasury to support growth and developer incentives.

Fee Burn Mechanism

Inspired by Ethereum’s EIP-1559, a configurable portion of each transaction fee can be burned to introduce deflationary pressure. Starts at 0%; adjustable via governance.

Supply & Inflation

2% annual inflation allocated to staking rewards—balancing sustainability with long-term participation.

ETH Restaking via EigenLayer

ETH holders can help secure Trunk by restaking through EigenLayer and earn bonus TRUNK rewards—bridging Ethereum capital with Solana-native infrastructure.

A Sustainable, Deflationary Economic Model. The TRUNK token powers the network’s staking, security, and governance layers

Sequencer

Rewards

Fee Burn

Supply & Inflation

ETH Restaking

Sequencer Rewards

70% of sequencer revenue goes to TRUNK stakers. The remaining 30% funds the ecosystem treasury to support growth and developer incentives.

Fee Burn Mechanism

Inspired by Ethereum’s EIP-1559, a configurable portion of each transaction fee can be burned to introduce deflationary pressure. Starts at 0%; adjustable via governance.

Supply & Inflation

2% annual inflation allocated to staking rewards—balancing sustainability with long-term participation.

ETH Restaking via EigenLayer

ETH holders can help secure Trunk by restaking through EigenLayer and earn bonus TRUNK rewards—bridging Ethereum capital with Solana-native infrastructure.

A Sustainable, Deflationary Economic Model. The TRUNK token powers the network’s staking, security, and governance layers

Sequencer

Rewards

Fee Burn

Supply & Inflation

ETH Restaking

Sequencer Rewards

70% of sequencer revenue goes to TRUNK stakers. The remaining 30% funds the ecosystem treasury to support growth and developer incentives.

Fee Burn Mechanism

Inspired by Ethereum’s EIP-1559, a configurable portion of each transaction fee can be burned to introduce deflationary pressure. Starts at 0%; adjustable via governance.

Supply & Inflation

2% annual inflation allocated to staking rewards—balancing sustainability with long-term participation.

ETH Restaking via EigenLayer

ETH holders can help secure Trunk by restaking through EigenLayer and earn bonus TRUNK rewards—bridging Ethereum capital with Solana-native infrastructure.

Trunk™ reduces the latency by over 65%*.

Read documentation

*Based on internal Trunk tests including 50 most popular MCPs and running over 10,000 unique tasks.

Launching the Future of Decentralized AI. Our Roadmap

Q2 2025

Root Phase: Testnet Launch (SVM + EigenDA)

Trunk begins in testnet with a Solana-native stack, powered by SVM execution and EigenDA for scalable data availability.

This phase validates Trunk’s core batching engine, MCP call routing, and stress-tests agent throughput under simulated load.

Q3 2025

Branch Phase: Mainnet Beta + Security Bounty

A production-grade beta goes live with a single sequencer to ensure tight feedback loops and operational stability.

A six-month, high-stakes security bounty will run in parallel to surface critical bugs ahead of opening up sequencing.

Q3 2025

Branch Phase: Mainnet Beta + Security Bounty

A production-grade beta goes live with a single sequencer to ensure tight feedback loops and operational stability.

A six-month, high-stakes security bounty will run in parallel to surface critical bugs ahead of opening up sequencing.

Q4 2025

Canopy Phase: Multi-Sequencer Auctions

Sequencing decentralizes. Multiple operators compete in bonded auctions to earn the right to batch and route agent requests—introducing price discovery and performance incentives to the protocol layer.

Q4 2025

Canopy Phase: Multi-Sequencer Auctions

Q2 2026

Trunk Phase: Permissionless Nodes + Agent-Prioritized Pricing

Trunk fully opens to the public: anyone meeting minimum technical thresholds can operate a node.

Pricing becomes dynamic, weighted by agent volume, task complexity, and MCP cost—aligning network incentives with real-world usage.

Q2 2026

Trunk Phase: Permissionless Nodes + Agent-Prioritized Pricing

Trunk fully opens to the public: anyone meeting minimum technical thresholds can operate a node.

Pricing becomes dynamic, weighted by agent volume, task complexity, and MCP cost—aligning network incentives with real-world usage.

Trunk — Specialized solana layer-2 optimized for AI workloads.

Social

Twitter

Product

Integrations

SOON

Enterprise

SOON

Resources