# Swarms

> Swarms is the multi-agent AI orchestration platform: an open-source Python framework, a hosted API with 15+ swarm architectures, an open marketplace for agents and prompts, and a free academy that teaches the entire stack. This site (swarms.ai) is the main company site; the marketplace runs at swarms.world, the API at api.swarms.world, and API docs at docs.swarms.ai.

## Main Pages

- [Products](https://www.swarms.ai/products): The complete multi-agent stack: Swarms Python, Swarms API, Swarms Marketplace, Swarms RS, Swarms Chat, and enterprise deployments.
- [The Swarms Stack](https://www.swarms.ai/stack): How the pieces fit: build agents with Swarms Python and Swarms Rust, deploy them with the Swarms API, monitor them with Swarms Cloud, and sell them on the Swarms Marketplace.
- [RouteHub](https://www.swarms.ai/routehub): The open-source LLM gateway behind Swarms: one completion() function for 20+ providers, built on the official OpenAI SDK, with LiteLLM-compatible names. Imports 383x faster than LiteLLM and adds 9 µs per call.
- [Swarms Framework](https://www.swarms.ai/framework): The production-grade Python framework for multi-agent orchestration.
- [Swarms API](https://www.swarms.ai/api): Hosted multi-agent orchestration with 1,000+ models and 15+ swarm architectures behind one endpoint.
- [Pricing](https://www.swarms.ai/pricing): Plans and pricing for the Swarms platform.
- [Installation](https://www.swarms.ai/installation): Install the Swarms framework and set up your environment.
- [Open Source](https://www.swarms.ai/open-source): The open-source repositories behind the Swarms ecosystem.
- [Research](https://www.swarms.ai/research): Research papers from the Swarms team: RouteHub and GraphWorkflow.
- [Mobile App](https://www.swarms.ai/mobile): Swarms for iOS and Android.
- [About Swarms](https://www.swarms.ai/about): How Swarms started: a 2023 research project putting multiple GPT-3 models in a group chat, which became the open-source Swarms framework and the full multi-agent stack.
- [The Swarms Story](https://www.swarms.ai/story): Timeline of every Swarms release with dates: Swarms Python (May 2023), Swarms Marketplace (Aug 2024), Swarms Rust (Mar 2025), Swarms API (May 2025), Swarms Cloud (Jun 2026), AgentHQ (Aug 2026), and RouteHub (Oct 2026).
- [Hiring](https://www.swarms.ai/hiring): Open roles at Swarms.

## Swarms Academy

- [Swarms Academy](https://www.swarms.ai/academy): Free courses for the multi-agent stack, from your first AI agent to production swarms.
- [The Swarms API Course](https://www.swarms.ai/academy/swarms-api): Free four-part multi-agent course with 31 lessons, 12 live API trials, checkpoint projects, and graded quizzes.
- [Part 1: Foundations: Your First Agent](https://www.swarms.ai/academy/swarms-api/foundations): Set up your environment and run your first agent completion.
- [Part 2: Single-Agent Capabilities](https://www.swarms.ai/academy/swarms-api/agent-capabilities): Memory, streaming, structured outputs, tools, MCP, and autonomy.
- [Part 3: Multi-Agent Orchestration](https://www.swarms.ai/academy/swarms-api/multi-agent-orchestration): Compose agents into swarms with the right architecture for the job.
- [Part 4: Production: Scale, Cost, and Operations](https://www.swarms.ai/academy/swarms-api/production): Batch execution, resilience, cost control, and observability.

## Blog

Every post is also available as raw markdown at /blog/<slug>/markdown (append /markdown to the post URL). Chinese translations live under /zh/blog/<slug>.

- [Swarms Launchpad Update: Now Built on Meteora DBC, With $SWARMS Pairs and 80% Faster Launches](https://www.swarms.ai/blog/swarms-launchpad-meteora-dbc): The Swarms Launchpad now launches tokenized agents and prompts directly on Meteora's Dynamic Bonding Curve. Pair your token with $SWARMS, earn creator fees in $SWARMS, keep earning after graduation, and launch about 80% faster. (中文: available)
- [Dynamic Model Pricing on the Swarms API: Every Model at Its Own Price](https://www.swarms.ai/blog/swarms-api-dynamic-model-pricing): The Swarms API no longer charges one flat token rate for every model. Each model is now billed at its provider's price plus a 15% service fee, the MCP and web scraper fees are gone, the search fee shows up in usage.total_cost, and a new /v1/models/pricing endpoint prices any model. Here is what changed, what it costs, and which endpoint outputs are different. (中文: available)
- [What Is Swarm Intelligence? From Ant Colonies to LLM Agent Swarms](https://www.swarms.ai/blog/what-is-swarm-intelligence): Swarm intelligence explained: stigmergy, boids, ant colony and particle swarm optimization, what LLM agent swarms borrow from them, plus Swarms API code. (中文: available)
- [What Is Multi-Agent Orchestration? Patterns, State, and Failure Handling](https://www.swarms.ai/blog/what-is-multi-agent-orchestration): Multi-agent orchestration explained: who runs when, what each agent sees, how outputs combine and when to stop, with patterns, failure handling and API code. (中文: available)
- [Swarm Architectures Explained: Every Way to Organize a Team of Agents](https://www.swarms.ai/blog/swarm-architectures-explained): A reference guide to every swarm architecture in the Swarms API: a diagram for each, when to use it, when to skip it, and the exact swarm_type payload to send. (中文: available)
- [Multi-Agent Collaboration Patterns: Debate, Voting, Mixture of Agents, and Councils](https://www.swarms.ai/blog/multi-agent-collaboration-patterns): Multi-agent collaboration patterns with working Swarms API code: debate, majority voting, Mixture of Agents, LLM councils, and the research behind each. (中文: available)
- [How to Build an Agent Swarm in Python: A Step-by-Step Tutorial](https://www.swarms.ai/blog/build-agent-swarm-python): Build an agent swarm in Python with the Swarms API: one example grows from a single agent to a pipeline, parallel specialists and a hierarchical swarm. (中文: available)
- [AutoGen Alternatives in 2026: Where to Go After AutoGen](https://www.swarms.ai/blog/autogen-alternatives): AutoGen is in maintenance mode. Compare each AutoGen alternative (Microsoft Agent Framework, AG2, Swarms, LangGraph, CrewAI) and see how to migrate your code. (中文: available)
- [Why Is LiteLLM Slow? Import Time, Per-Call Overhead and How to Fix It](https://www.swarms.ai/blog/why-is-litellm-slow): Why is LiteLLM slow? We measure its 1.2 second import, per-call overhead, streaming cost and memory, then cover the documented fixes and when to switch. (中文: available)
- [How to Migrate from LiteLLM to RouteHub: A Step-by-Step Guide](https://www.swarms.ai/blog/migrate-from-litellm-to-routehub): Migrate from LiteLLM to RouteHub step by step: swap imports, move litellm globals to per-call arguments, update responses, tools, errors and tests. (中文: available)
- [Introducing RouteHub: A Lightning-Fast LLM Gateway](https://www.swarms.ai/blog/introducing-routehub): RouteHub is an open-source LLM gateway that connects 20+ providers through one API, with streaming, async and tool calling. It imports 383x faster than LiteLLM, reaches its first response 5.5x sooner with 3.9x less memory, adds 9 µs per call, and is faster than the official Anthropic SDK on Claude calls and streaming. This post covers how it works, the full benchmark results and how to get started. (中文: available)
- [How to Use Claude with the OpenAI SDK in Python Without Losing Caching](https://www.swarms.ai/blog/claude-openai-format-python): Learn how to use Claude with the OpenAI SDK format in Python, what Anthropic's compatibility endpoint drops, and how to keep prompt caching, thinking and PDFs. (中文: available)
- [Best LLM Gateway for Python in 2026: RouteHub, LiteLLM, any-llm, aisuite](https://www.swarms.ai/blog/best-llm-gateway-python): Looking for the best LLM gateway for Python? We compare RouteHub, LiteLLM, any-llm and aisuite on startup time, per-call overhead, streaming and features. (中文: available)
- [Best LiteLLM Alternative in 2026: RouteHub and 5 Other Options](https://www.swarms.ai/blog/best-litellm-alternative): Looking for a LiteLLM alternative? We compare RouteHub, any-llm, aisuite, Portkey, Bifrost and OpenRouter on speed, dependencies, features and license. (中文: available)
- [Swarms Is Now on Docker Hub: How to Run Agents and Swarms in Docker](https://www.swarms.ai/blog/swarms-docker-hub-official-image): The official Swarms image, swarmscorp/swarms, is live on Docker Hub with Python 3.13, the swarms package and the swarms CLI for amd64 and arm64. Learn how to pull it, run your first agent, pass API keys safely, give agents tools, run multi-agent workflows, use Docker Compose and package your own agent as an image. (中文: available)
- [Get Verified on the Swarms Marketplace](https://www.swarms.ai/blog/get-verified-swarms-marketplace): Pro and Premium members now get a blue Verified badge next to their username on swarms.world. Here is what the badge means, how it turns on and off, what each plan includes, and why creators who sell on the marketplace should get it. (中文: available)
- [Expanded MCP Support on the Swarms Marketplace: Inspect, Connect, and Publish MCP Servers](https://www.swarms.ai/blog/expanded-mcp-support-swarms-marketplace): Publishing an MCP server on swarms.world now takes three steps: paste the URL, inspect it, publish. The marketplace connects to your server, reads its tools, and detects how it authenticates. Every MCP listing then gets a connect section with ready-to-paste configs for Claude, Cursor, VS Code, and other MCP clients. Here is how each piece works and how to publish your own server. (中文: available)
- [What Is a Decision Model? How Jev and Clef Differ from LLMs](https://www.swarms.ai/blog/what-is-a-decision-model): What is a decision model? How TypeSafe Jev and Cloudflare Clef answer typed questions with calibrated confidence instead of text, and how to use them in Swarms. (中文: available)
- [How to Turn an AI Agent into an MCP Server in Python (with Auth)](https://www.swarms.ai/blog/turn-ai-agent-into-mcp-server-python): Turn an AI agent into an MCP server in Python with Swarms MCPDeployer: API keys, custom auth, token verifiers, swarms as tools, and a client agent to call it. (中文: available)
- [Tree of Thoughts in Python: A Practical Implementation Guide](https://www.swarms.ai/blog/tree-of-thoughts-python-guide): Learn tree of thoughts in Python with Swarms: solve the Game of 24, compare BFS and DFS, propose vs sample, value vs vote, cap cost, and read real call counts. (中文: available)
- [How to Track LLM Token Usage and Cost per Agent in Python](https://www.swarms.ai/blog/track-llm-token-usage-cost-python): Track LLM token usage and cost per agent in Python: input, output, cached and reasoning tokens, per-run snapshots, dollar costs and swarm-wide totals in Swarms. (中文: available)
- [Swarms x TypeSafe: How to Use Decision Models with the Swarms Framework and API](https://www.swarms.ai/blog/swarms-typesafe-decision-models): A complete guide to decision models in Swarms. Ask TypeSafe's Jev and Cloudflare's Clef typed questions for calibrated probabilities, compare and ensemble several models, plug them into agents and swarms, and call them through the Swarms API or the TypeSafe SDK. (中文: available)
- [Reduce LLM Costs 10x: Screen with a Decision Model First](https://www.swarms.ai/blog/reduce-llm-costs-decision-model): Reduce LLM costs by screening every item with a decision model and sending only the top 5% to LLM agents. Runnable Swarms code, measured costs and timings. (中文: available)
- [Multi-Agent System Failure Modes: 7 Silent Failures and Fixes](https://www.swarms.ai/blog/multi-agent-system-failure-modes): Seven multi-agent system failure modes that return plausible output instead of an error, how to detect each one, and the fixes in Swarms 16, with runnable code. (中文: available)
- [Swarms v16 'Overclock': Token Accounting, Decision Models, MCP Deployment, and Runs That Start Clean](https://www.swarms.ai/blog/swarms-v16-overclock-release): The complete technical changelog for Swarms v16, code-named Overclock. Every agent and swarm now reports what it cost, a new DecisionModel brings typed, calibrated decisions from TypeSafe and Cloudflare into any workflow, MCPDeployer serves agents as authenticated MCP servers, TreeOfThoughts lands, a dozen structures stop leaking one task's conversation into the next, failed LLM calls finally raise, and tool handling moves out of agent.py into a ToolManager that makes a tool turn up to 65x faster. Every new feature, improvement, and bug fix from September 1 to October 2, 2026, day by day. (中文: available)
- [Every Prompt on the Swarms Marketplace Now Has a Security Score](https://www.swarms.ai/blog/prompt-security-scores-swarms-marketplace): Prompt pages on swarms.world now carry a Security tab and a 0 to 100 Security Score, powered by SkillScanner. Here is what the score means, how the scan works, who can see it, and how to use it (and SkillScanner itself) to check a prompt before you run it or publish it. (中文: available)
- [Swarms Rust Benchmarks: 6 ms Cold Starts, 3.7 MB of Memory, and 100 Parallel Agents in 0.52 Seconds](https://www.swarms.ai/blog/swarms-rust-benchmarks): We benchmarked swarms-rs 0.3.0 against Swarms (Python), LangGraph and CrewAI on Claude Sonnet 5.5. swarms-rs starts in 6 ms (130x to 440x faster), idles in 3.7 MB (25x to 68x less memory), adds 0.11 ms of framework time per LLM call (10x to 88x less), and runs 100 agents in parallel in 0.52 seconds against an ideal of 0.50. This post covers the method, every result with charts, and how to reproduce the numbers yourself. (中文: available)
- [Swarms Rust v0.3.0: OpenRouter, Any Model by Name, Sub-Agents and Handoffs](https://www.swarms.ai/blog/swarms-rust-v0-3-0-changelog): swarms-rs 0.3.0 and swarms-macro 0.2.0 are on crates.io: an OpenRouter provider, AnyModel for choosing any provider with one string, sub-agents, handoffs, typed tool outputs, a router that works with every provider, working support for current Claude models, and fixes for deadlocks, silent failures and broken tool schemas. (中文: available)
- [Introducing SkillScanner: Security Auditing for AI Agent Skills and Prompts](https://www.swarms.ai/blog/introducing-skillscanner): SkillScanner is an open source security scanner for AI agent skills and prompts. It combines 49 static detection rules with a Swarms agent review to catch prompt injection, malicious links, credential theft, hidden instructions, and supply chain risk before a skill reaches your agents. Scan a folder, a URL, or raw text from Python, a REST API, or Docker. (中文: available)
- [Swarms Weekly Ecosystem Update [September 20 - September 27]: A Faster Marketplace, Private GitHub Imports, Quick Launch, and Five New Guides](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-sep-20-27): This week across the Swarms ecosystem: more than 100 performance changes landed on the Swarms Marketplace, with a home page that shows listings 3.1× faster and creator profiles carrying 44% less code; private GitHub repositories can now be imported straight into a listing; the new Quick Launch page puts a tokenized agent live from one short form; Swarms Chat gained full agent controls and every model the API reports; and five new guides explain Swarms Cloud and compare Swarms with CrewAI, the OpenAI Agents SDK, AutoGen and LangGraph. (中文: available)
- [The Swarms Marketplace Is Faster: September Performance Update](https://www.swarms.ai/blog/marketplace-performance-update-september-2026): Today we are shipping a major speed update to the Swarms Marketplace. The home page shows listings 3.1× faster, a 494 KB download is gone from every page, and listing pages carry up to 44% less code. Live now at swarms.world. (中文: available)
- [What Is Swarms Cloud? The Control Plane for Multi-Agent Systems](https://www.swarms.ai/blog/what-is-swarms-cloud): A complete guide to Swarms Cloud: one API key across 2,000+ models and 16 swarm architectures, visual and generated agent builders, Batch and Grid runners that scale a single task to thousands of completions, a page for every agent and every completion, an encrypted Skills library, a hosted MCP server, and scale-to-zero agent hosting with S2A. (中文: available)
- [Swarms vs CrewAI: Which Multi-Agent Framework Should You Use?](https://www.swarms.ai/blog/swarms-vs-crewai): A framework-level comparison with working code: the same pipeline in CrewAI and in Swarms, then the capabilities Swarms has that CrewAI does not. Topology as a parameter across 15+ structures, a compiled graph engine, dynamic tool loading that sends one schema instead of fifty, a 16-tool autonomous harness, first-class MCP, per-agent model choice across providers, and typed chat turns that keep multi-agent context correct. (中文: available)
- [OpenAI Agents SDK vs Swarms: The Complete 2026 Comparison](https://www.swarms.ai/blog/openai-agents-sdk-vs-swarms): A detailed comparison of OpenAI's Agents SDK and Swarms: model access, the OpenAI-compatible endpoint, handoffs, 16 orchestration architectures, cross-vendor failover, pricing, tools, MCP, streaming, and observability. Runnable code for both, plus real run times and costs from the live Swarms API. (中文: available)
- [Swarms vs AutoGen vs LangGraph vs CrewAI: The 2026 Multi-Agent Framework Comparison](https://www.swarms.ai/blog/multi-agent-framework-comparison-2026): Four multi-agent frameworks, one task in each, and the only published reproducible orchestration benchmark in the category: Swarms runs compiled agent graphs at a 7.0x geometric mean over LangGraph, up to 62.5x on deep chains. Why the compile-once design wins, and who it does not fit. (中文: available)
- [Migrating from LangGraph to Swarms: A Working Guide](https://www.swarms.ai/blog/migrating-from-langgraph-to-swarms): Fourteen graph patterns built twice, in LangGraph and in Swarms GraphWorkflow, each with a diagram and runnable code: pipelines, fan-out, fan-in, diamonds, layered meshes, map-reduce, multi-vendor juries, debates, planner and workers, unrolled reflection, gated routing, subgraphs, and edge metadata. Plus how GraphWorkflow compiles and runs a graph, a feature-by-feature comparison, costs, and the benchmark case: 7.0x faster execution on average and up to 62.5x on deep chains. (中文: available)
- [Import Private GitHub Repositories Into the Swarms Marketplace](https://www.swarms.ai/blog/import-private-github-repos): You can now publish an agent straight from a private repository. Connect GitHub once, choose exactly which repositories to share, and import without making any of your code public. Read-only access, one-hour tokens, nothing stored, revocable from GitHub at any time. (中文: available)
- [Swarms Weekly Ecosystem Update [September 13 - September 20]: Completion Logs Ship, the Cloud Changelog Lands, and The Prompt Royale Closes](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-sep-13-20): This week across the Swarms ecosystem: completion logs went live on Swarms Cloud with a page for every API request, the August 3 to September 19 changelog covered seven weeks of platform work including Auto Agent Builder, Chat, Batch, Grid, Skills and MCP, The Prompt Royale ran its full seven days with $5,000 in SOL on the line, the Marketplace MCP server got its deep dive, and one Swarms account now signs you into every product in the stack. (中文: available)
- [Everything New in Swarms Chat](https://www.swarms.ai/blog/swarms-chat-updates): Swarms Chat has been rebuilt over the past several weeks: conversation starters that arrive in under three seconds, a settings panel exposing every agent parameter, a searchable picker covering every model the API reports, per-reply token counts, retry and continue, a full mobile pass, and a set of fixes for bugs that were quietly dropping data. (中文: available)
- [Swarms Cloud Changelog: August 3 to September 19](https://www.swarms.ai/blog/swarms-cloud-changelog-august-september): Seven weeks of platform work, day by day: the Auto Agent Builder, multi-agent Chat, Batch runs, the Grid runner, an encrypted Skills library, the MCP page, an Examples gallery, a page per agent and a page per completion, S2A deployments, syntax highlighting everywhere, a dashboard rebuilt for speed, and a long run of billing, security and correctness fixes. (中文: available)
- [Introducing Completion Logs on Swarms Cloud](https://www.swarms.ai/blog/completion-logs-swarms-cloud): Every API request your account makes is now tracked in one place. Search and filter by agent, endpoint, or ID, monitor token usage and cost per run, and open any completion on its own page to inspect the full request payload and response. (中文: available)
- [The Swarms Marketplace MCP Is Live: 6,000+ Agents, Prompts, and Tools Behind One Connection](https://www.swarms.ai/blog/swarms-marketplace-mcp-server): swarms.world is now a Model Context Protocol server. Connect any MCP client to swarms.world/mcp and your agent can search 6,000+ agents, prompts, and tools, read free listings in full, publish and edit its own, leave reviews, and check creator fees. Here is what it does, how it is built, and how to connect in under a minute. (中文: available)
- [A Better Swarms Chat Experience Is Here](https://www.swarms.ai/blog/swarms-chat-improvements): We redesigned the chat experience on the Swarms Marketplace. Every prompt now opens in a full-screen chat, with a settings panel that exposes the full agent spec, a cleaner header, and a model picker that lists nearly 2,000 models from the Swarms API with Claude 5, GPT-5.4+, GLM 5, and Kimi K3 pinned to the top. (中文: available)
- [How to Build Agents and Multi-Agent Systems with GPT-6 Astra in Swarms](https://www.swarms.ai/blog/getting-started-gpt-6-astra-swarms): A step-by-step tutorial for OpenAI's GPT-6 Astra: build a custom agent and a Mixture of Agents with the open-source Swarms framework, then run the same thing through the Swarms API with no infrastructure, per-run cost reporting, and 50% off token costs at night. (中文: available)
- [Swarms Weekly Ecosystem Update [August 31 - September 6]: Swarms v15 Akira Ships, the USDC Competition Closes at $5M, and New MCP Tutorials](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-aug-31-sep-6): This week across the Swarms ecosystem: Swarms v15 Akira shipped with dynamic tool loading and a rebuilt autonomous agent harness, the USDC Competition closed with nearly $5M in trading volume and 3,500+ traders, the Marketplace passed 6,125 products from 720+ sellers, a new set of MCP tutorials covered DeepWiki, Exa, Firecrawl, Hugging Face, and Semgrep, and Swarms published guidance for staying online through a multi-provider frontier model outage. (中文: available)
- [Fee Cards Are Live: Turn Your Creator Earnings Into a Shareable Card](https://www.swarms.ai/blog/fee-cards-share-creator-earnings): Every tokenized agent and prompt on the Swarms Marketplace pays its creator a share of trading fees. Fee Cards turn that number into a shareable image with your avatar, handle, ticker, and total fees earned in USD. Here is what they show, how to find yours, and why posting it is the best marketing your agent can get. (中文: available)
- [Unique Agent Configuration Pages Are Live on Swarms Cloud](https://www.swarms.ai/blog/agent-detail-pages-swarms-cloud): Swarms Cloud gives each agent in your account a page of its own: activity and cost, the full system prompt, all 34 configuration fields with their live defaults, and a per-run view of how much of the model's context window each run actually filled. (中文: available)
- [Swarms v15 'Akira': Dynamic Tool Loading, a Real Agent Harness, and Typed Chat Turns](https://www.swarms.ai/blog/swarms-v15-akira-release): The complete technical changelog for Swarms v15, code-named Akira. A DynamicToolLoader defers tool schemas behind a searchable catalog so a fifty-tool agent sends one schema, the autonomous agent harness ships sixteen working built-in tools with an honest transcript and a mutable plan, every multi-agent structure now delivers context as typed chat turns instead of one flattened blob, and 27 dead modules are gone. Every new feature, improvement, and bug fix from July 29 to September 1, 2026, day by day. (中文: available)
- [Swarms Weekly Ecosystem Update [August 24 - August 31]: AgentHQ Pre-Beta, Swarms v15 Akira, and the Builder's Guide to Tokenized Agents](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-august-24-31): This week across the Swarms ecosystem: AgentHQ entered pre-beta as an interactive orchestration environment for Claude and Codex agents, Swarms v15 Akira was announced with up to 40 percent lower token usage across multi-agent architectures, the Builder's Guide to Tokenized AI Agents was published, and the USDC Launch Championship opened on Swarms Launchpad. (中文: available)
- [Introducing AgentHQ: The Office Where Your AI Agents Actually Work](https://www.swarms.ai/blog/introducing-agenthq): AgentHQ turns multi-agent orchestration into a place you can walk around. Hire real Claude and Codex agents, give them desks, assign work, and watch every tool call happen live. Here is the architecture behind it, the four ways agents can talk to each other, the office themes, and how to get early access. AgentHQ is in pre-beta: join the waitlist at swarms.ai/agenthq. (中文: available)
- [The USDC Launch Championship: Price Your Agent in Dollars](https://www.swarms.ai/blog/usdc-launch-championship): The new Swarms Launchpad competition is live. Launch an agent or prompt with USDC as the quote currency between August 26 and September 2 UTC and compete for a share of a $1,000 USDC prize pool, ranked by total trading volume. (中文: available)
- [The Complete Guide to Tokenizing Your Agent](https://www.swarms.ai/blog/complete-guide-tokenizing-your-agent): Why an AI agent is a productive asset, what tokenization solves for builders in attention, ownership, and revenue, why people invest in tokenized agents, and every decision in the Swarms Launchpad flow from connection type to Frenzy Mode, USDC, and Vault Mode. (中文: available)
- [Swarms Weekly Ecosystem Update [August 17 - August 24]: The Swarms MCP Server, MCP Scribe, and Encrypted Skills Storage](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-august-17-24): This week across the Swarms ecosystem: the hosted Swarms MCP server went live with 23 tools behind a single endpoint, the MCP Portal shipped on Swarms Cloud with live status and connection diagnostics, MCP Scribe turned OpenAPI specifications into production MCP servers, Swarms Cloud added encrypted prompt and skills storage with native SKILL.md support, and the Marketplace changelog covered more than 200 improvements shipped between July 21 and August 21. (中文: available)
- [The Swarms MCP Server Is Live: One Endpoint, Twenty-Three Tools, No Local Install](https://www.swarms.ai/blog/mcp-server-live-swarms-cloud): Swarms Cloud now has a dedicated MCP page at cloud.swarms.world/mcp for the hosted Model Context Protocol server at mcp.swarms.world/mcp. Point any MCP client at one URL to give it agents, swarms, batch execution, and account telemetry as callable tools. The page carries live endpoint status, connection snippets for Python and TypeScript, swarm completion examples, and client tutorials. (中文: available)
- [Swarms Marketplace Changelog July 21st - August 21st: Competitions, the Screener, MCP Servers, and a Public API](https://www.swarms.ai/blog/swarms-marketplace-changelog-july-august): A day by day log of everything that shipped on the Swarms Marketplace over the last month: on-chain agent competitions with real prize pools, the tokenized agent Screener, a dedicated MCP Servers page, a self-updating public API spec, a rebuilt home page and sign-in, a full mobile pass, and a serious security hardening effort, each written in plain language so you know what changed and what it means for you. (中文: available)
- [GraphWorkflow: Our New Research Paper on a Compile-Once Engine That Runs Agent Graphs up to 62.5x Faster Than LangGraph](https://www.swarms.ai/blog/graphworkflow-research-paper): The Swarms research team has published a full systems paper on GraphWorkflow, the graph execution engine inside the Swarms framework. Across an open benchmark suite of five topologies at 10 to 200 nodes, GraphWorkflow executes compiled agent graphs with a geometric-mean speedup of 7.0x over LangGraph, rising to 62.5x on deep chains, compiles graphs 21.6x to 31.3x faster, and completes the cold build-compile-execute path 7.9x faster. This article walks through the paper: the cost taxonomy, the compile-once architecture, the programming model comparison, the full benchmark results with figures from the paper, and how to reproduce every number yourself. (中文: available)
- [Skills: Ultra-Secure Private Prompt and Skill Storage, Now Live on Swarms Cloud](https://www.swarms.ai/blog/private-skills-encrypted-storage-swarms-cloud): Swarms Cloud now has a private, encrypted library for your prompts and skills at cloud.swarms.world/skills. Save prompts by hand or drag and drop Anthropic-format SKILL.md files, organize them with tags and search, give every entry its own page, and store all of it encrypted with a key derived from your account, so only you can ever read it. Available to every user on every plan: Free, Pro, and Premium. (中文: available)
- [Swarms Weekly Ecosystem Update [August 11 - August 16]: The Tokenized Agent Screener, the GMGN Integration, and New Cloud Tutorials](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-august-11-16): This week across the Swarms ecosystem: the Screener puts every tokenized agent on one live page with on-chain market data, the GMGN integration makes Swarms agents discoverable and tradable on one of Solana's largest trading platforms, two new video tutorials cover the Auto Agent Builder and the Compare feature on Swarms Cloud, the Vault Mode Competition enters its final days, off-peak pricing cuts API token costs in half overnight, and Swarms opens hiring across engineering, research, growth, finance, and operations. (中文: available)
- [Learn How to Build Multi-Agent Systems With Swarms Cloud: Two New Video Tutorials](https://www.swarms.ai/blog/swarms-cloud-tutorials): Two new video tutorials cover the Auto Agent Builder and the Compare feature on Swarms Cloud: generate complete AI agent teams from natural language, and run the same task through multiple agents to evaluate their outputs side by side. A full Swarms Cloud course is coming soon. (中文: available)
- [Swarms x GMGN: Discover and Trade Swarms Agents on GMGN](https://www.swarms.ai/blog/swarms-gmgn-integration): Swarms, the largest AI launchpad on Solana, is now integrated with GMGN. Filter Swarms agents in the Trenches tab, find any agent by contract address, and buy tokenized agents through GMGN in one click from the Swarms Launchpad. (中文: available)
- [Introducing the Screener: Every Tokenized Agent on the Swarms Marketplace, Live on One Page](https://www.swarms.ai/blog/tokenized-agents-screener): The Swarms Marketplace now has a dedicated Screener for tokenized agents and prompts. Watch trending agent tokens ranked by market cap, filter by Frenzy Mode and Vault Mode, and see new launches pop onto the page in near real-time the moment they go live. (中文: available)
- [Swarms Weekly Ecosystem Update [August 3 - August 8]: The Auto Agent Builder, the MCP Portal, and a Two-Month Cloud Changelog](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-august-3-8): This week across the Swarms ecosystem: the Auto Agent Builder generates a complete agent team from a single task description, the MCP Portal opens a discovery and publishing surface for Model Context Protocol servers, a changelog documents seven weeks of Swarms Cloud reliability work, the Marketplace adds direct address routing, and the first Marketplace competition closes as the next one opens. (中文: available)
- [Introducing the Auto Agent Builder: Generate a Complete Agent Team From a Single Task](https://www.swarms.ai/blog/introducing-auto-agent-builder): The Auto Agent Builder turns a one-line task into a full roster of ready-to-run agent configurations, each with its own name, system prompt, and model. Here is how it works, how to use the page and the API, and how to take a generated roster straight into production. (中文: available)
- [Introducing the MCP Portal: Discover, Publish, and Monetize MCP Servers on the Swarms Marketplace](https://www.swarms.ai/blog/mcp-portal-swarms-marketplace): The Swarms Marketplace now has a dedicated MCP Portal for Model Context Protocol servers. Browse the most popular community-built MCPs, connect tools like Firecrawl, DeepWiki, and Exa to your AI agents, and publish your own MCP server with free or paid access. (中文: available)
- [The Vaulted Tokenized Agents Competition Is Live: $2,000 in SOL Over Two Weeks](https://www.swarms.ai/blog/vaulted-tokenized-agents-competition): The next Swarms Marketplace competition is here. Launch an AI agent with Vault Mode enabled before August 19 and compete for a $2,000 SOL prize pool: $1,000 for first, $700 for second, $300 for third, ranked by real on-chain trading volume. (中文: available)
- [The First Frenzy Competition Results, and the Vaulted Tokenized Agents Competition](https://www.swarms.ai/blog/frenzy-competition-results-vault-mode-competition): The first Swarms Frenzy Competition closed with 23+ agent launches and over $250,000 in trading volume. Here are the winners, how prizes are paid, and the next competition: two weeks of Vault Mode launches competing for $2,000 in SOL. (中文: available)
- [Swarms Cloud Changelog: July 14 to August 3](https://www.swarms.ai/blog/swarms-cloud-changelog-july-august): A week by week, day by day log of everything that shipped on Swarms Cloud over the last three weeks: live swarm architecture pages fetched straight from the API, a public pricing page, a session persistence fix, a security hardening pass, saved flows in the Workflow Builder, a rebuilt Token Usage date range, and a footer connection check. (中文: available)
- [Swarms Weekly Ecosystem Update [July 27 - August 1]: Swarms v14 'Zena', Agent Competitions, and the API Examples Suite](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-july-27-aug-1): This week across the Swarms ecosystem: Swarms v14 'Zena' ships AutoAgentBuilder, a unified MCP manager, and a faster GraphWorkflow; the Marketplace launches recurring Agent Competitions; a new open-source repository collects 100+ runnable Swarms API examples; and two new guides cover GraphWorkflow versus LangGraph and building agents with Kimi K3. (中文: available)
- [Swarms v14 'Zena': AutoAgentBuilder, a Unified MCP Manager, and a Faster GraphWorkflow](https://www.swarms.ai/blog/swarms-v14-zena-release): A complete technical walkthrough of Swarms v14, code-named Zena. AutoAgentBuilder turns a plain-English task into a working agent roster, a unified MCP manager brings OAuth and multi-server tool discovery, and GraphWorkflow now runs on native rustworkx. Plus three new multi-agent architectures, sandboxed computer-use tools, distributed tracing, and a security pass covering SSRF and token storage. Every feature, improvement, and bug fix since June 12, with runnable code. (中文: available)
- [Ship AI Agents Faster with the Swarms API Examples Suite](https://www.swarms.ai/blog/swarms-api-examples-suite): Our new open-source repository collects 100+ runnable examples for the Swarms API, covering single agents, coordinated swarms, every orchestration pattern, model integrations across GPT-5, Claude, Gemini, Grok, DeepSeek and more, plus streaming, vision, tools, and real industry use cases. (中文: available)
- [Introducing Agent Competitions: Build, Launch, and Compete for Real Prizes](https://www.swarms.ai/blog/agent-competitions-swarms-marketplace): Swarms now runs recurring competitions for agent builders. The first one, the Frenzy Volume Championship, pays $1,000 in SOL to the three agents with the most trading volume this week. Entry is automatic, the leaderboard is live, and a new competition follows every week. (中文: available)
- [Swarms GraphWorkflow vs LangGraph: A Simpler, Faster Way to Build Agent Graphs](https://www.swarms.ai/blog/swarms-graphworkflow-vs-langgraph): A hands-on comparison of Swarms GraphWorkflow and LangGraph for building directed-acyclic-graph agent pipelines. Same task, both frameworks, real code, showing why Swarms is easier to use, parallel by default, and faster to ship. (中文: available)
- [How to Get Started with Kimi K3 in Swarms: A Step-by-Step Tutorial](https://www.swarms.ai/blog/getting-started-kimi-k3-swarms): A beginner-friendly guide to building AI agents with Moonshot's Kimi K3 model in the Swarms framework. Install with uv, run a single agent, then orchestrate a multi-agent GroupChat of Kimi agents in minutes. (中文: available)
- [Swarms Weekly Ecosystem Update [July 21-26]: The Frenzy Hub, a New Marketplace Home Screen, and Gemini 3.6 Flash on Swarms Cloud](https://www.swarms.ai/blog/swarms-weekly-ecosystem-update-july-21-26): This week across the Swarms ecosystem: the Frenzy Hub brings live market data for every tokenized agent into one page, the Marketplace ships a new home screen, a changelog documents close to 200 Marketplace updates in seven weeks, and Gemini 3.6 Flash arrives on Swarms Cloud with a full developer guide. (中文: available)
- [The Frenzy Hub: A Live Dashboard for Every Agent Token on Swarms](https://www.swarms.ai/blog/frenzy-hub-live-agent-token-dashboard): Swarms now has a dedicated home for Frenzy Mode agent tokens: live market caps that re-quote every 15 seconds, a creator leaderboard, a filterable screener across every frenzy launch, and a launch panel that carries your settings straight into publishing. (中文: available)
- [Who We Are: Swarms Corporation and the Mission to Build Collective Superintelligence](https://www.swarms.ai/blog/who-is-swarms-corporation): Swarms Corporation exists to build Collective Superintelligence: we research multi-agent systems and deploy them into the real economy. This is who we are, what we build, and why we believe the future of intelligence is a network of agents, not a single mind. (中文: available)
- [Swarms Marketplace Changelog June 1st - July 21: Fiat Payments, Upvotes, and More](https://www.swarms.ai/blog/swarms-marketplace-changelog-june-july): A day by day log of everything that shipped on the Swarms Marketplace from June through July: marketplace bundles, an upvote system, card payments through Stripe, natural language search, faster pages, unique avatars, and dozens of fixes, each written in plain language so you know what changed and what it means for you. (中文: available)
- [Building Agents on Swarms with Gemini 3.6 Flash: A Developer Guide](https://www.swarms.ai/blog/swarms-gemini-3-6-flash): Google's new Gemini 3.6 Flash and 3.5 Flash-Lite models are available on the Swarms API. This step-by-step guide shows developers how to get an API key, run a single agent, and orchestrate a multi-agent swarm using the new Flash series, with full Python and curl examples. (中文: available)
- [Swarms Weekly Recap [July 13-20]: Fiat Payments on Swarms Marketplace, 7,000 GitHub Stars, Swarms Academy](https://www.swarms.ai/blog/swarms-weekly-roundup-july-2026): A roundup of what shipped across the Swarms ecosystem this week: the launch of Swarms Academy, 7,000 GitHub stars, live fiat payments on the Marketplace, a two-month Swarms Cloud changelog, and a founders keynote on the autonomous economy. (中文: available)
- [The Micro Agents Running the Swarms Marketplace](https://www.swarms.ai/blog/specialized-micro-agents-marketplace): Tags, tickers, images, summaries, search, and support on the Swarms Marketplace are now handled by small specialized agents that do one job each. Here is every one of them, how they compose, and why the marketplace gets easier to use every week. (中文: available)
- [Vault Mode: How to Turn a Tokenized Agent Into a Holder-Gated Product](https://www.swarms.ai/blog/vault-mode-tokenized-agents): A deep dive into Vault Mode on the Swarms Marketplace: exactly how the holder gate works end to end, how it interacts with pricing, how to launch a vaulted agent or prompt, and why creators use it to turn usage demand into token buy pressure. (中文: available)
- [Every User Now Gets a Unique Agent Avatar](https://www.swarms.ai/blog/unique-agent-avatars): Users who haven't uploaded a profile picture no longer get a gray circle or a plain initial. Swarms now generates a unique, deterministic agent avatar for every one of them, computed instantly in the browser with zero network requests. (中文: available)
- [Tokenized Agents: What They Are, How to Launch One, and How Creators Earn Ongoing Revenue](https://www.swarms.ai/blog/tokenized-agents-launch-revenue): A complete guide to tokenizing an AI agent or prompt on the Swarms Marketplace: how the Solana bonding curve works, the exact fee split between creator and platform, Frenzy Mode, Vault Mode, and how to launch your first tokenized agent. (中文: available)
- [How to Sell Your Agents and Prompts on the Swarms Marketplace: A Step-by-Step Guide to Stripe and Crypto Payouts](https://www.swarms.ai/blog/how-to-sell-agents-prompts-stripe-crypto): A practical, step-by-step guide to selling AI agents, prompts, and tools on the Swarms Marketplace. Set up Stripe seller payouts, price your listing, choose a payment rail, and understand exactly how fiat and crypto payouts work. (中文: available)
- [How to Find Agents, Prompts, Tools, and Bundles on the Swarms Registry](https://www.swarms.ai/blog/how-to-find-agents-registry-page): A practical walkthrough of the Swarms Registry page: the search bar, AI auto-search, every filter, sort order, grid vs. table view, and the refresh button, so you can find the right agent or prompt in seconds. (中文: available)
- [Frenzy Mode: How to Launch a Tokenized Agent With 2x Creator Fees](https://www.swarms.ai/blog/frenzy-mode-tokenized-agents): A focused deep dive into Frenzy Mode on the Swarms Marketplace: the exact fee mechanics, how Frenzy-mode tokens show up in the registry, and a step-by-step guide to launching a tokenized agent or prompt with it enabled. (中文: available)
- [Swarms Reaches 7,000 GitHub Stars](https://www.swarms.ai/blog/swarms-7000-github-stars): Four years after it began as a research project, the open-source Swarms framework has crossed 7,000 GitHub stars, with 961 forks and 58 contributors. A thank you to the community that built it, and a look at where the project goes from here. (中文: available)
- [Introducing Swarms Academy: Free Courses for the Multi-Agent Stack](https://www.swarms.ai/blog/introducing-swarms-academy): Swarms Academy is live at swarms.ai/academy. The flagship Swarms API course spans four parts and 31 lessons, with live API trials you run directly in the page, checkpoint projects, graded quizzes, and a points system that ranks you from Recruit to Swarm Architect. All of it is free. (中文: available)
- [The Agent Economy Now Takes Card, Powered by Stripe](https://www.swarms.ai/blog/fiat-payments-live-swarms-marketplace): The Swarms Marketplace now supports fiat payments powered by Stripe. Buyers in 100+ countries can purchase AI agents and prompts with cards, Apple Pay, Google Pay, Alipay, Klarna, and more, while sellers get automatic bank payouts on every sale. (中文: available)
- [Swarms Cloud Changelog: Two Months of Platform Updates](https://www.swarms.ai/blog/swarms-cloud-changelog): A week by week, day by day log of everything that shipped on Swarms Cloud over the past two months: a redesigned Agents workspace, a public model catalog spanning 1,500+ models, subscription billing, free credit for new accounts, guided onboarding, and dozens of reliability fixes, with what each change actually means for you. (中文: available)
- [What Is Collective Superintelligence?](https://www.swarms.ai/blog/what-is-collective-superintelligence): A clear definition of Collective Superintelligence (CSI): what it is, the components it is built from, how it differs from AGI and ASI, how a swarm of specialized agents actually produces intelligence beyond any single model, and how to start building one today. (中文: available)
- [What Is an AI Agent? LLM, Tools, Memory, and a Loop](https://www.swarms.ai/blog/what-is-an-agent): A precise definition of an AI agent: a large language model wrapped in tools, memory, and a looping structure. What each component does, why the loop is what makes an agent autonomous, and how to build your first agent in a few lines of Python. (中文: available)
- [What Is an Agent Marketplace? How the Swarms Marketplace Lets You Discover and Trade AI Agents and Prompts](https://www.swarms.ai/blog/what-is-an-agent-marketplace): A complete guide to agent marketplaces: what they are, why they exist, and how the Swarms Marketplace at swarms.world lets you discover, publish, trade, and monetize AI agents, prompts, and tools, with pricing, tokenization, and code examples. (中文: available)
- [Collective Superintelligence: Why CSI Will Surpass AGI and ASI](https://www.swarms.ai/blog/collective-superintelligence): Collective Superintelligence (CSI), networks of specialized agents working in concert, will surpass AGI in intelligence and capability while creating far greater economic value. Here is the argument, axis by axis: parallelism, reasoning depth, fault tolerance, memory, and specialization. (中文: available)
- [Introducing Autonomous Filter Search: Describe What You Want, an Agent Finds It](https://www.swarms.ai/blog/autonomous-filter-search-agent): The Swarms Marketplace registry now ships with an autonomous filter search agent. Describe the agents, prompts, or tools you are looking for in plain English, and an AI agent translates your request into precise registry filters, applies them in front of you, and even recovers when nothing matches. (中文: available)
- [The Autonomous Economy Is Coming: Our Founders Keynote at AI Salon](https://www.swarms.ai/blog/autonomous-economy-keynote): Swarms co-founder and CEO Kye Gomez addressed more than 80 attendees at AI Salon on the future of the economy: collapsing cognitive labor costs, a second labor force of billions of agents, and the rise of corporations that run themselves. Includes a full quickstart for the Swarms API. (中文: available)
- [Speeding Up the Agent Economy: Major Marketplace Performance Improvements](https://www.swarms.ai/blog/marketplace-speed-improvements): The Swarms Marketplace now loads over 60% faster, with dramatically quicker product pages, lower-latency registry search, and a more responsive browsing experience across 6,000+ agents and prompts. (中文: available)
- [Inside the Swarms Cloud Workflow Builder: Visual Multi-Agent Graphs](https://www.swarms.ai/blog/workflow-builder-swarms-cloud): A detailed look at how the Swarms Cloud Workflow Builder turns agents into a visual, connectable graph, how the underlying graph workflow execution model works, and how to build a production multi-agent pipeline with it. (中文: available)
- [What Is a Multi-Agent System?](https://www.swarms.ai/blog/what-is-a-multi-agent-system): A clear, no-fluff explanation of what a multi-agent system actually is, why single LLM calls hit a ceiling, and the handful of ways multiple AI agents coordinate to get real work done. (中文: available)
- [Single Agent vs. Multi-Agent: When One Agent Isn't Enough](https://www.swarms.ai/blog/single-agent-vs-multi-agent): A practical decision framework for choosing between a single well-tooled agent and a multi-agent architecture, with concrete signals, a worked example, and a checklist. (中文: available)
- [Manager-Worker Agent Architectures: A Practical Introduction](https://www.swarms.ai/blog/manager-worker-agent-architectures): How director agents plan, delegate, and aggregate work across specialist worker agents, and when the extra layer of hierarchy is actually worth it. (中文: available)
- [Compare Agents Side by Side on Swarms Cloud](https://www.swarms.ai/blog/compare-agents-swarms-cloud): Swarms Cloud's new Compare tool lets teams run the same task through two to six agent configurations at once, then see real cost, token usage, and output differences side by side before committing to production. (中文: available)
- [Agentic AI Explained: How Autonomous Agents Actually Decide What to Do](https://www.swarms.ai/blog/agentic-ai-explained): A precise, mechanics-first explanation of agentic AI: the perceive-reason-act loop, how agents call tools mid-task, how memory shapes decisions, and why reasoning strategy is a configurable knob rather than a fixed property of the model. (中文: available)
- [Sequential, Concurrent, and Graph Workflows: The Three Shapes of Agent Orchestration](https://www.swarms.ai/blog/agent-orchestration-patterns): Every multi-agent pipeline reduces to one of three topologies: a line, a fan-out, or a DAG. A framework-agnostic look at sequential, concurrent, and graph orchestration, with real code for each. (中文: available)
- [All-New Swarms Cloud Updates](https://www.swarms.ai/blog/all-new-swarms-cloud-updates): Swarms Cloud adds a token usage dashboard, model pages spanning 1,500+ models, subscription management, and a wave of platform improvements for teams building production multi-agent systems. (中文: available)
- [Swarms v2.0: The First Multi-Agent Framework in Rust](https://www.swarms.ai/blog/swarms-v2-rust-framework): Introducing the revolutionary Rust-based multi-agent framework that delivers unprecedented performance and memory safety for enterprise AI deployments. (中文: available)
- [Advanced Agent-to-Agent Communication Protocols](https://www.swarms.ai/blog/agent-communication-protocols): Deep dive into the sophisticated communication protocols that enable seamless collaboration between AI agents in complex workflows. (中文: available)
- [Enterprise Deployment Guide: Scaling AI Agents in Production](https://www.swarms.ai/blog/enterprise-deployment-guide): Comprehensive guide to deploying and scaling multi-agent systems in enterprise environments with maximum reliability and performance. (中文: available)

## Ecosystem

- [Swarms API Docs](https://docs.swarms.ai): Full API reference, guides, and examples. Serves its own llms.txt at https://docs.swarms.ai/llms.txt.
- [Swarms Framework Docs](https://docs.swarms.world): Python framework documentation and quickstart.
- [Swarms Marketplace](https://swarms.world): Discover, buy, and sell AI agents, prompts, and tools.
- [GitHub](https://github.com/kyegomez/swarms): The open-source Swarms framework repository.