Everything you need to deploy, run, and monitor AI in production.
- Agent-as-a-ServiceCreate your agents, then embed them in your product or internal tools with one API.
- Workflows and automationAutomate any process on a visual canvas: triggers, branches, approvals, and integrations.
- Monitoring and securityEvery agent and workflow run, visible, isolated, and under control.
Designed to move from experimentation to production with just a few clicks.
Ship agents inside your product
Build agents in the dashboard. Chat, stream, and manage them visually or through one REST API.
Read the docs
OpenClaw
Hermes
Pi Agent
LangGraph
Anthropic
OpenAI
- Visual builder to configure and deploy sandboxed agents.
- REST API with sync and streaming (SSE) chat.
- Programmatic lifecycle: configure, create, start, stop, and reload via API.
- Non-ephemeral storage for agent workspaces.
- Master keys for your account, scoped keys per agent.
Automate with or without agents
A drag-and-drop canvas for triggers, branches, approvals, and integrations. Run it standalone or wire it to your agents.
Launch Console- Manual, webhook, schedule, and polling triggers.
- Start workflows when an agent starts, or call agents from a run.
- HTTP, code, and integrations like Gmail, Slack, and Notion in one canvas.
- Human approval steps inside a run.
Monitoring and security, built in
Usage and spend dashboards for your agents, step-level run history for your workflows. Isolated sandboxes and encrypted credentials.
Talk to us
- Usage, spend, and tokens per agent and per tenant.
- Step-level run history for every workflow.
- Isolated sandboxes, one per agent.
- Live runtime logs from inside each sandbox.
Ship AI products, don't fight infrastructure.
JUST COMPARE
See how Guayaba stacks up against the alternatives, from self-hosted setups to general-purpose AI tools.
| Feature |  | (DIY) |
|---|---|---|
| Open-source harness support | ✓ | ✓ |
| Time to first launch | Minutes | Hours to days |
| Guided setup | ✓ | ✕ |
| No servers or DevOps needed | ✓ | ✕ |
| Real use-case templates | ✓ | ✕ |
| Workflow automation | ✓ | Manual |
| Agent management from one place | ✓ | ✕ |
| Agent visibility and observability | ✓ | Manual |
| Agent-as-a-Service via API | ✓ | ✕ |
| Best for | Teams running agents and workflows in production without ops work | Self-hosters and power users |
| Main limitation | Less raw control than full DIY | High setup and maintenance burden |
Simple plans for running AI in production
Choose the plan that matches how many agents you want to run and how many credits you need for usage and compute.
Best for trying Guayaba with a single small agent.
- ✕Multiple agent instances
- ✕Unlimited running time
- ✕Priority Support
- ✓OpenAI Free Models
- ✓Basic Dashboard
- ✓2000 welcome credits
For the first teams shipping agents with us.
- ✕Multiple agent instances
- ✓Unlimited running time
- ✕Priority Support
- ✓OpenAI Free Models
- ✓Basic Dashboard
- ✓3000 monthly credits
For teams running several agents in production.
- ✓Up to 3 agent instances
- ✓Unlimited running time
- ✓Priority Support
- ✓OpenAI Free Models
- ✓Basic Dashboard
- ✓10000 monthly credits
For organizations that need private deployments, stricter compliance requirements, and more control over infrastructure.
- ✓Unlimited agent instances
- ✓Unlimited running time
- ✓Priority Support
- ✓Custom model
- ✓Basic Dashboard
- ✓Custom monthly credits
- ✓Custom hardware configuration
- ✓Private deployment options
- ✓Stricter compliance support
- ✓Advanced platform features
- ✓Dedicated account manager
- ✓SLA-backed support
Frequently Asked Questions
Not to get started. Configure agents through the dashboard and guided setup. Developers can plug into the API when they need deeper control.
Common starting points include customer support agents, lead qualification flows, research assistants, and internal copilots. If your use case is more specific, we can help you shape a custom template for it.
Yes. You can launch and manage agents from the dashboard, then connect them to your product, workflow, or internal tools through the API.
Yes. Guayaba includes a visual workflow builder with manual, webhook, schedule, and polling triggers, human approval steps, and integrations. Workflows run on the same platform as your agents.
A GRAFT is a reusable agent template that bundles agent configuration, tools, and behavior into a deployable starting point. It helps you get to production faster.
Yes. Each agent runs as an independent instance with isolated resources, and you can run OpenClaw and Hermes agents side by side from the same dashboard.
Choose from the managed model catalog, or connect your own model provider accounts with BYOK.
Agents run in isolated, tenant-scoped environments with encrypted credentials. If your requirements are stricter, talk to us and we can walk through your setup.
Yes. If you already know the workflow you want to automate but are not sure how to structure the agent, talk to us on Telegram or email us at david@guayaba.run.
Pricing depends on your plan and available credits. You pay for platform access plus the usage and compute resources your agents consume.
Your managed AI cloud
for agents and workflows.
Deploy and manage AI agents and workflows without handling servers or operational overhead yourself.
Launch through the dashboard. Integrate through the API.
