VoltAgent
An open-source TypeScript framework and enterprise platform for building, observing, and deploying AI agents with multi-agent orchestration, RAG, and workflow automation.
At a Glance
Pricing
Perfect for getting started. Free forever with limited usage.
Engagement
Available On
Listed Mar 2026
About VoltAgent
VoltAgent is an end-to-end AI agent engineering platform that combines an open-source TypeScript framework with the VoltOps cloud console for observability, automation, and deployment. Built by the team behind Refine, it brings clarity and best practices to JavaScript AI development — filling a gap left by Python-dominated tooling. The platform supports everything from single-agent prototypes to enterprise-scale multi-agent systems, with full TypeScript type safety and Zod schema validation throughout.
- Open-Source Core Framework: Install via
npm create voltagent-app@latestto scaffold a production-ready TypeScript agent with built-in support for Memory, RAG, Guardrails, Tools, MCP, Voice, and Workflow. - Multi-Agent Orchestration: Define a Supervisor Agent with
subAgentsto coordinate specialized agents, each with their own LLM, tools, and memory, enabling complex reasoning pipelines. - Workflow Chain API: Use
createWorkflowChain()with.andAll()andandAgent()to compose declarative, fully-typed multi-step agent workflows with pause/resume support for human-in-the-loop scenarios. - VoltOps Observability Console: Connect your agents to the VoltOps console to monitor LLM calls, trace agent behavior, run evals, manage prompts, and debug workflows in real time — available as cloud or self-hosted.
- RAG Integration: Use
.embed(),.query(), and.rerank()APIs to build knowledge bases backed by vector stores like Pinecone, Postgres, and Supabase, with support for OpenAI, Cohere, and Voyage embeddings. - Triggers & Actions: Configure event-driven triggers (e.g., Slack messages, webhooks) to automate agent workflows across 40+ integrations including GitHub, Gmail, Salesforce, Jira, and Stripe.
- Unified LLM Provider API: Switch between OpenAI, Anthropic, and custom LLMs with a single line change, using the
@ai-sdkadapter pattern. - Prompt Management: Version and iterate prompts directly in the VoltOps console without redeploying code, enabling rapid experimentation and A/B testing.
- Deployment: Deploy agents to production with persistent tunnel URLs, basic auth, and self-hosted options available on higher-tier plans.
Community Discussions
Be the first to start a conversation about VoltAgent
Share your experience with VoltAgent, ask questions, or help others learn from your insights.
Pricing
Free Plan Available
Perfect for getting started. Free forever with limited usage.
- 1 seat
- 1 prompt
- 1 project
- 250 traces/month
- 7 days data retention
Core
Perfect for small teams and production use.
- Up to 3 seats
- Unlimited prompts & agents
- Included runtime: 43,200 min
- 50,000 traces/month + Pay-as-you-scale
- 90 days retention
- 5,000 Actions/month
- 5,000 Triggers/month
- 5,000 Evals/month
- Persistent tunnel URLs
- RAG: 50 knowledge bases, 10 documents/knowledge base
- Uploads: 10 files/batch, max 100 MB each
- RAG requests: 100/min
- 14-day free trial
Pro
Ideal for larger teams with advanced needs.
- Up to 20 seats
- Included runtime: 129,600 min
- 250,000 traces/month + Pay-as-you-scale
- 20,000 Actions/month
- 20,000 Triggers/month
- 20,000 Evals/month
- Deployment basic auth
- SSO Integration
- Priority Support
- RAG: Unlimited knowledge bases, 20 documents/knowledge base
- Uploads: 20 files/batch, max 100 MB each
- RAG requests: 1,000/min
- 14-day free trial
Enterprise
For large-scale deployments with dedicated support and self-hosted options.
- Dedicated account manager
- Self-hosted deployment
- Unlimited tracing
- Custom MSA
- Slack Priority Support
- Feature prioritization
- Unlimited seats
- Unlimited prompts & agents
- Unlimited data retention
Capabilities
Key Features
- Open-source TypeScript agent framework
- Multi-agent supervisor orchestration
- Workflow Chain API with pause/resume
- LLM observability and tracing
- RAG with vector database integration
- Triggers and actions for event-driven automation
- Prompt management and versioning
- Agent evals and evaluation pipelines
- Persistent memory across conversations
- MCP (Model Context Protocol) support
- Voice agent support
- Guardrails for agent safety
- Tool calling and function execution
- Unified LLM provider API
- Self-hosted deployment option
- SSO integration (Pro/Enterprise)
- 40+ app integrations
