Everything you need to build, deploy, and scale AI agent workflows with SwarmAI.
Build your first AI swarm in 5 minutes. Agents, tasks, processes, and tools.
Define workflows declaratively in YAML. Zero Java code required.
Multi-tenancy, governance gates, budget tracking, RBAC, and audit trail.
Configure OpenAI, Anthropic, Ollama, and other LLM provider API keys.
Run SwarmAI in Docker with Redis, PostgreSQL, and ChromaDB.
SwarmAI is organized into 8 focused Maven modules
swarmai-coreAgents, tasks, processes, state, skills, memory, knowledge, budget, governance, observability
swarmai-tools46 built-in tools: web search, PDF, CSV, shell, HTTP, email, SEC filings, CVE/OSV lookup, GitHub PR
swarmai-dslYAML DSL parser & compiler for declarative workflow definitions
swarmai-rlLightweight RL: contextual bandits, Thompson sampling, heuristic policies
swarmai-enterpriseMulti-tenancy, deep RL (DQN), advanced governance, RBAC, audit, billing
swarmai-evalSelf-evaluation swarm, competitive benchmarks, value scoring, regression detection
swarmai-studioWeb dashboard for workflow monitoring, event replay, and debugging
swarmai-bomBill of Materials for dependency version alignment
<!-- Add to your pom.xml -->
<dependency>
<groupId>ai.intelliswarm</groupId>
<artifactId>swarmai-core</artifactId>
<version>1.0.28</version>
</dependency>How to adopt SwarmAI in an existing Java application — what changes, what stays, and how to move an existing agent workflow across.
Configure model-provider and tool API keys for SwarmAI — which keys each provider and built-in tool requires, and where to set them.
Run a complete agentic workflow in Docker with local Ollama models — no cloud API key required.
Design proposal for federated policy improvement: how SwarmAI deployments contribute learned improvements without sharing workflow content.
Skill sandbox, prompt-injection defence, memory retrieval and policy correctness in SwarmAI 1.0.29.
What changed in SwarmAI 1.0.24.
What changed in SwarmAI 1.0.23 — Docker-backed tools.