Summoner provides SPLT, the first protocol designed specifically for AI agents to securely communicate, coordinate, and exchange value across environments. By creating this essential foundation layer, we're transforming isolated AI capabilities into powerful collaborative networks that bridge legacy and agentic systems while enabling entirely new categories of autonomous economies.
Connect. Secure. Transact.
Today's AI agents are powerful but isolated—unable to communicate, coordinate, or exchange value in standardized ways. Summoner solves this fundamental challenge with SPLT (Summoner Protocol for Logic Transport), the first protocol specifically designed for autonomous agent networks. Just as HTTP enabled the web, SPLT provides the missing foundation for the autonomous economy, allowing developers to build agent systems that seamlessly communicate across environments, bridge web2 and web3, and form dynamic networks with built-in security and economic mechanisms. By standardizing how agents interact while maintaining flexibility, Summoner is creating the infrastructure where the true potential of collaborative AI can finally be realized.
Discover Our Features
Summoner is the next generation foundational protocol layer for autonomous agent networks.
The Summoner Agent SDK empowers developers to build, deploy, and coordinate multiple autonomous agents with programmable automations, and flexible token-based rate limits.
Programmable workflows using Tickers, periodic transactions executed automatically, and Embeds, triggerable transactions fired by external events or other agents.
We created a flexible server architecture that can either run a Rust or Python backend, coordinated through a Python wrapper. The Python-based server is designed for experimentation and testing, while the Rust-based server is designed for high performance.
Our protocol has enterprise-grade security built with finite state machine architecture for predictable agent behavior. Our default communication has APIs with extensible custom methods and built-in economic incentives for agentic services.
We created a proprietary evaluation system where the rewards are the combination of individual agent's quality of execution (QoE) and agent group's success (COQ, Collective Outcome Quality).
Our differentiator is that each agent individually has its own SPLT protocol, but also collectively, the agent group is also part of the SPLT protocol, enabling many agents-to-many agents communication.
Discover how Summoner is transforming industries through autonomous agent networks
In the near future, agents will negotiate with each other in break-neck speeds to solve unprecedented complex problems. Summoner provides a framework for an automated negotiation protocol between two autonomous agents: SellerAgent and BuyerAgent.
Our protocol demonstrates asynchronous communication, dynamic pricing strategies, randomized behaviors, and real-time tracking of negotiation outcomes. In this system, two agents engage in continous price negotiations:
The buyer tries to purchase at the lowest possible price, adjusting upward if needed.
The seller tries to sell at the highest possible price, adjusting downward if needed.
A negotiation succeeds when both agents agree on a price.
Otherwise, the negotiation fails and a new session begins.
Modern customer support must scale across countless channels, languages, and support topics—all while delivering fast, accurate, and empathetic responses. Traditional chatbot systems struggle with edge cases and siloed understanding, and individual AI agents often lack the context or collaboration needed for end-to-end resolution.
Summoner enables a cooperative network of specialized customer support agents that dynamically collaborate to understand customer intent, resolve issues, and continuously learn from interactions to improve performance.
Intent Detection & Routing Agent: Identifies the customer’s core issue from messages, voice, or ticket data, and routes it to the optimal set of specialist agents.
Knowledge Base Retrieval Agent: Queries internal documentation, past tickets, and public help articles to surface accurate, context-aware responses.
Sentiment & Escalation Agent: Analyzes tone, urgency, and frustration levels in real time to adjust responses or trigger escalation to human support if needed.
Post-Resolution Feedback Agent: Follows up with customers, gathers feedback, and relays insights to improve agent performance and update support content.
Financial markets move at incredible speeds with countless variables affecting asset prices. Human traders cannot process the volume of signals required for optimal decision-making, while individual AI models often lack the comprehensive perspective needed for confident execution.
Summoner enables a collaborative network of specialized trading agents working in unison to analyze market data, reach consensus, and execute trades with unprecedented efficiency.
Market sentiment analysis across news feeds, social media, and commentary
Technical analysis processing price patterns and chart formations
Fundamental analysis evaluating financials and economic indicators
Risk management and optimized order execution
Meet the founders behind Summoner
Kevin is a seasoned technology leader with extensive experience in technical product and operations roles at Consensys–MetaMask, Mozilla Firefox, and various startup media companies. As an entrepreneur, he led MomsCorner—a social platform for parents—from inception to a successful public listing, now trading under the ticker BHLD.
Elliott is an infrastructure engineer and designer. He started his career in 2017 contributing to QRL. His code executes over 250 billion times per second at Meta, across over a million physical machines, saving millions of dollars in energy per year.
Rémy is a mathematician and systems designer focused on cryptographic protocols, multi-agent architectures, and complex system modeling. He completed postdoctoral research at MIT after training at École Normale Supérieure. His work combines formal methods, logic, and principles from biology to design secure, self-organizing, and privacy-preserving computational systems.
Get in touch to learn more about Summoner
Whether you're an enterprise looking to leverage autonomous agent networks, a developer interested in building on our platform, or an investor exploring new opportunities, we'd love to hear from you.
