Cryptonian🩺 (Swarms)

Swarms Marketplace is extending the reach of AI agents through its dedicated MCP Server Marketplace. The platform brings together a growing collection of MCP servers, from web search and data extraction to databases, APIs, coding and research, users can discover and connect tools that enable agents to interact with external services. Developers can also bring their own MCP servers to the marketplace, making their tools accessible to a wider audience. Each listing comes with its capabilities and access details, allowing users to evaluate and integrate available tools into their workflows. With support for free and paid listings, builders can distribute their MCP servers and generate revenue from their work.

Swarms Launchpad trading activity, happening live and recorded on-chain. Each transaction reflects actual market movement, with tokens being traded, liquidity circulating and volume accumulating across the ecosystem. The live feed captures these developments as they happen, offering a clear view of individual trades and overall market activity across launched tokens. Powered by Meteora DBC, Swarms brings token launches and subsequent trading into a connected on-chain environment. From individual exchanges to wider market movements, the entire process remains transparent and visible.

An existing API can become a tool that AI agents can discover and use through MCP. Swarms has published a practical guide showing developers how to turn any HTTP API into an MCP server and list it on the Marketplace. Using Python and FastMCP, the tutorial demonstrates how to: • Convert API endpoints into callable MCP tools. • Define typed arguments and tool descriptions. • Test the server and implement bearer-token authentication. • Deploy it through Docker with a public HTTPS endpoint. • Publish it on Swarms Marketplace. Once listed, Swarms automatically inspects the server, reads its available tools and generates connection instructions for clients like Claude, Cursor and VS Code. Developers can offer free or paid access, with 90% of every sale going to the creator.

Swarms Framework v17 introduces Comms Channels, a major step forward in multi-agent coordination. Agents can now communicate through shared channels, publishing messages, accessing previous contributions and maintaining context across separate executions. The feature supports six implementations: File, Queue, SQLite, HTTP, SMTP and AgentMail. This provides multiple ways to coordinate agents, whether through local storage, databases, message queues, remote servers or email. With persistent communication and shared context, agents can build on one another's work instead of repeatedly starting from scratch. Comms Channels make more coordinated, connected and persistent multi-agent workflows possible.

Swarms Framework v17, codenamed InterLink, is here. This release brings a significant improvement to how agents communicate, coordinate and execute tasks. Key developments include shared Comms Channels, Phased Fan-Out Workflows, improved concurrency and RouteHub integration. Framework import time has dropped from 1.91 seconds to just 0.28 seconds. Concurrent and graph workflows can now execute entire layers of agents without the previous default concurrency restrictions. The release also introduces improved streaming, cleaner execution states and more reliable error handling. With 227 commits across 271 files, v17 makes Swarms faster, more connected and considerably leaner.