Directory
Apps
Every page, tool, and product in Swarms Cloud, grouped by purpose. Press ⌘K to jump anywhere.
Real-time analytics for your Swarms workspace: agent counts, executions, success rate, and rate-limit windows across minute/hour/day.
API request log explorer backed by /v1/account/logs. Search and paginate every agent and swarm run with token usage and cost breakdowns.
Every event on your account in one searchable timeline: agent and swarm completions, API keys created, credit purchases, auto-recharges and redeemed credit codes. Filter by type and date, and export to CSV.
Charts of input/output tokens and cost across your completions, by day, week, or month, grouped by agent or API key, with filters, a table view, and CSV export.
Paste or upload up to 500 tasks, pick an agent configuration, and fan them out through the batch completions endpoint in chunks of 50. Results stream into a table with per-task status, output, tokens, and cost, exportable as CSV.
Send up to 50 tasks and 50 agents through the batched grid workflow endpoint in one call, then read the results as a matrix: tasks are rows, agents are columns. Click any cell for the full output, and export the whole grid as CSV.
Author single agents (model, prompt, temperature, tools) and execute them against the production Swarms API in one click.
Catalog of agent configurations from /v1/agents/list. Search, sort, and inspect every persisted agent on your account.
Interactive composer for multi-agent swarms. Pick agents, pick an architecture, send a task, and inspect the output.
Build a roster of agents and hold an ongoing conversation with it. One agent chats directly with full memory of the conversation; add more and every agent answers each turn with the transcript as context.
Describe a task and a builder agent designs the team for you — complete AgentSpec configurations with names, prompts, and models you can post unmodified to the swarm endpoints. Backed by /v1/auto-agent-builder/completions.
Configure or import 2 to 6 agents, send them the same task, and compare their responses, latency, tokens, and cost side by side.
Drag agents onto a canvas, connect them into a directed flow, then run the workflow on the platform or export the exact /v1/graph-workflow/completions request as Python, TypeScript, Go, or cURL. Pro/Ultra plans.
A playground for /v1/decision/completions and SystemOne (/v1/systemone): ask noul, choice and score questions about any text and get calibrated probabilities back, with live model prices, the question types, the endpoints and code in cURL, Python and TypeScript.
A private library for your prompts and skills. Write them here or import Anthropic-format SKILL.md files, markdown, or JSON by drag and drop; tag, search, and reuse them in any agent. Everything is encrypted and visible only to you, and each entry gets its own page.
Deploy agents and services from a GitHub repository, a Dockerfile, or a pre-built image. S2A builds the image, picks the cheapest region across AWS, Google Cloud and Azure, gates the rollout on a healthcheck, and serves it behind a stable URL that scales to zero when idle. In beta: deploy from the Deployments page.
Live model catalog from /v1/models/available: GPT, Claude, Gemini, Llama and more. Cached for 10 hours; refresh on demand.
Every architecture from /v1/swarms/available with descriptions: Hierarchical, Sequential, Concurrent, Mixture of Agents, Council, Debate, Router, Auto Builder, and more.
Every reasoning agent on /v1/reasoning-agent/completions: Self-Consistency for majority-vote answers, Reflexion for self-critique and refinement, and Agent Judge for evaluating outputs. Architecture diagrams, use cases and quick starts.
Build with the Swarms API in any language. Official, typed SDKs for Python, TypeScript, Go, Java, and C#, plus the MCP server for Claude Desktop and Cursor.
A hosted Model Context Protocol server for the Swarms API at api.swarms.world/mcp. Live endpoint status and observed uptime, connection snippets for Python and TypeScript, swarm completion examples, and client tutorials.
The Swarms API in US East, US West, Europe and Asia Pacific, each with its own URL and live health. Keep inference in the United States or in Europe with a regional URL, with Python, TypeScript and cURL examples.
A specialized Prompt Architect agent (claude-sonnet-4-6) turns a one-line brief into a drop-in deployable system prompt with named sections and operating constraints.
Free, Pro and Premium side by side: which endpoints and models each plan reaches, rate limits, compute lane and support. Each model is billed at its own token price, the same on every plan, so a subscription buys access and throughput rather than cheaper tokens.
Estimate API costs by tokens, agents, images and Exa search, plus night discount and Frenzy Mode.
Per-minute, per-hour, and per-day windows with live usage bars, countdowns to each reset, tier details, and upgrade options.
Black Friday and Cyber Monday offers: 100% off tokens for Pro and Premium on Black Friday and Cyber Monday, $15 in starting credits for new accounts, 50% off the first 3 months of a monthly plan, an extra 10% off the first year of an annual plan, 20% bonus credits on top-ups.
Manage the API keys that authenticate your apps against the Swarms API. Full keys are shown once at creation; revoke compromised keys instantly.
Manage your Swarms API key, view your credit balance (paid, free, referral), and choose the workspace theme.
All billing on your account in one table: subscription invoices, credit purchases, auto-recharges and redeemed credit codes, with Stripe receipt links, totals, search, filters and CSV export.
Eight worked calls against the Swarms API — single agents, agents with images, concurrent and sequential swarms, hierarchical swarms, batch completions, and account calls — each with a runnable request in cURL, Python, TypeScript, Go, and Rust.
The Swarms Cloud in one page: capabilities, all 16 swarm architectures, the management console, global infrastructure, pricing, and SDKs.