Built for real use cases
and real integrations
Guayaba helps you move from a concrete workflow to a live AI agent faster. Start from templates, configure through a guided dashboard, and connect your agent to your product or internal tools through the API.
Use-case templates
Start from real workflows instead of a blank page. Use curated GRAFT templates for support, research, lead capture, internal operations, and more.
Guided deployment
Configure your agent, connect your keys, and launch. Guayaba handles provisioning, runtime setup, and the operational overhead behind the scenes.
Agent-as-a-Service API
Launch through the dashboard, then connect your AI agent to your app, workflow, or internal tools through the API.
Designed to move from experimentation to production with just a few clicks.
Launch ConsoleOpenClaw-native deployment
Guayaba is built to make OpenClaw easier to configure, launch, and manage in production through a guided dashboard.
The fastest path from idea to a live OpenClaw agent.
Agent-as-a-Service API
Launch and manage OpenClaw agents from the dashboard, then expose them inside your product, workflow, or internal tools through the API.
Use Guayaba as the backend layer for agent-powered products.
Verified templates (GRAFTs)
A GRAFT (Guayaba Runtime Agent Framework Template) is a reusable starting point that bundles agent behavior, tools, and runtime configuration into a deployable template.
Start from a real template instead of building every setup from scratch.
Common use cases
Launch templates for customer support, lead qualification, research assistants, internal copilots, and more.
Start from the job you want done, not from the infrastructure.
Production-grade security
Credentials and configurations are protected with modern security controls, encrypted communications, and account safeguards.
Built for teams that need reliability, access control, and trust.
Air-Gapped Inference
Deploy models in isolated environments. Prompts, responses, and model weights never touch external servers. Zero data egress. Enterprise compliance for SOC2, GDPR, HIPAA.
Private inference by design. Your data never leaves your infrastructure.
Observability and monitoring
Track your agent status, usage, and runtime from one place so you always know what is live and how it is performing.
More visibility, less guesswork.
Launch in 1 minute, Deliver as a Service
Read the docsJUST COMPARE
See how Guayaba stacks up against the alternatives, from self-hosted setups to general-purpose AI tools.
| Feature | |  (DIY) |
|---|---|---|
| Built for OpenClaw | ✓ | ✓ |
| Time to first launch | Minutes | Hours to days |
| Guided setup | ✓ | ✕ |
| No servers / DevOps needed | ✓ | ✕ |
| Real use-case templates | ✓ | ✕ |
| Help shaping the right agent for your workflow | ✓ | ✕ |
| Runtime visibility | ✓ | Manual |
| Observability into what the agent is doing | ✓ | Manual |
| Agent management from one place | ✓ | ✕ |
| OpenClaw ready for real workflows | ✓ | Yes, with manual setup |
| Agent-as-a-Service via API | ✓ | ✕ |
| Good for teams | ✓ | Technical teams only |
| Good for non-technical users | ✓ | ✕ |
| Best for | Teams and users who want a real OpenClaw workflow without setup pain | Self-hosters and power users |
| Main limitation | Less raw control than full DIY | High setup and maintenance burden |
| Feature | | (DIY) |  |  | MaxClaw | AtomicBot |
|---|---|---|---|---|---|---|
| Built for OpenClaw | ✓ | ✓ | ✕ | ✕ | ✕ | ✓ |
| Time to first launch | Minutes | Hours to days | Minutes | Days to weeks | Minutes | Minutes |
| Guided setup | ✓ | ✕ | ✓ | ✕ | ✓ | ✓ |
| No servers / DevOps needed | ✓ | ✕ | ✓ | Usually no | ✓ | ✓ |
| Real use-case templates | ✓ | ✕ | ✕ | ✕ | Limited | Limited |
| Help shaping the right agent for your workflow | ✓ | ✕ | ✕ | ✕ | ✕ | ✕ |
| Runtime visibility | ✓ | Manual | Limited | Partial | Limited | Limited |
| Observability into what the agent is doing | ✓ | Manual | Limited | Partial | Limited | Limited |
| Agent management from one place | ✓ | ✕ | Partial | Partial | Partial | Partial |
| OpenClaw ready for real workflows | ✓ | Yes, with manual setup | ✕ | ✕ | ✕ | Partial |
| Agent-as-a-Service via API | ✓ | ✕ | ✕ | Partial, but not OpenClaw AaaS | ✕ | ✕ |
| Good for teams | ✓ | Technical teams only | Partial | Engineering teams | Partial | Limited |
| Good for non-technical users | ✓ | ✕ | ✓ | ✕ | ✓ | ✓ |
| Best for | Teams and users who want a real OpenClaw workflow without setup pain | Self-hosters and power users | General AI assistance | Code-first teams building their own stack | Hosted assistant experience | Faster personal OpenClaw setup |
| Main limitation | Less raw control than full DIY | High setup and maintenance burden | Not OpenClaw-first | Requires engineering work | Less control and less OpenClaw workflow depth | Fast start, but limited observability, limited template depth, and no Agent-as-a-Service layer |
Frequently Asked Questions
Not to get started. You can configure an agent through the dashboard and guided setup flow. Developers can still 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 OpenClaw agents from the dashboard, then connect them to your product, workflow, or internal tools through the API.
A GRAFT is a reusable agent template that bundles runtime 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, so you can operate more than one agent from the same dashboard.
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.
For privacy-sensitive setups, Guayaba supports isolated deployment options so prompts, responses, and model weights stay under tighter control.
It can be a strong fit for teams that need stricter deployment boundaries, private model options, and more control over data handling.
Pricing depends on your plan and available credits. You pay for platform access plus the usage and compute resources your agents consume.
Simple plans for getting AI agents into 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 experimenting and launching a single small agent.
- ✕Multiple agent instances
- ✕Unlimited running time
- ✕Priority Support
- ✓OpenAI Free Models
- ✓Basic Dashboard
- ✓2000 welcome credits
Current early access plan for first users.
- ✕Multiple agent instances
- ✓Unlimited running time
- ✕Priority Support
- ✓OpenAI Free Models
- ✓Basic Dashboard
- ✓3000 monthly credits
For advanced users and small teams running agents more consistently.
- ✓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
Your control plane
for AI agents
Configure, deploy, and manage AI agents without handling servers or operational overhead yourself.
Launch through the dashboard. Integrate through the API.
