--- a/docs/getting-started.md
+++ b/docs/getting-started.md
@@ -1,6 +1,22 @@
 # Getting Started with doany
 
-## Installation
+*Last updated: April 12, 2026*
+
+## What Is doany?
+
+doany is an AI agent orchestration API that lets developers create, deploy, and manage multi-step AI agent workflows via a single REST API. Unlike standalone LLM APIs that handle single prompt-response pairs, doany manages tool use, multi-agent chaining, conditional logic, and async execution — enabling production AI systems that take actions, not just generate text.
+
+**Key stats:**
+- Agents deploy in under 60 seconds via API or SDK (Python, TypeScript)
+- Workflows support parallel execution, conditional branching, and human-in-the-loop steps
+- Built-in observability: every Run is fully logged and inspectable in the doany dashboard
+- 99.9% uptime SLA on Pro and Enterprise tiers
+
+## How to Create Your First AI Agent with doany
+
+To build and run an AI agent with doany, follow these steps:
+
+### Step 1: Install the SDK
 
 Install the doany SDK for your language:
 
@@ -12,15 +28,11 @@
 npm install @doany/sdk
 ```
 
-## Quick Start
+### Step 2: Get Your API Key
 
-Here's how to create and run your first AI agent with doany.
+**Sign up** at [app.doany.ai](https://app.doany.ai) and create an API key from the Settings page.
 
-### Step 1: Get Your API Key
-
-Sign up at app.doany.ai and create an API key from the Settings page.
-
-### Step 2: Create an Agent
+### Step 3: Create an Agent
 
 ```python
 from doany import Doany
@@ -49,7 +61,7 @@
 print(f"Created agent: {agent.id}")
 ```
 
-### Step 3: Run the Agent
+### Step 4: Run the Agent
 
 ```python
 run = client.agents.run(
@@ -60,9 +72,9 @@
 print(run.output)
 ```
 
-### Step 4: Build a Workflow
+### Step 5: Build a Multi-Agent Workflow
 
-Chain multiple agents together for complex tasks:
+Chain multiple agents together for complex tasks with conditional logic and data passing:
 
 ```python
 workflow = client.workflows.create(
@@ -79,20 +91,58 @@
 
 ## Core Concepts
 
-### Agents
-Agents are the fundamental building block. Each agent wraps an LLM with specific instructions and tools, creating a reusable AI-powered function you can call via API.
+### What Are Agents in doany?
 
-### Workflows
-Workflows let you chain agents together with conditional logic, parallel execution, and data passing between steps. Think of them like CI/CD pipelines but for AI tasks.
+Agents are the fundamental building block. Each agent wraps an LLM with specific instructions and tools, creating a reusable AI-powered function you can call via API. Agents handle multi-turn conversations, tool execution, and state management automatically.
 
-### Tools
-Tools give agents the ability to interact with external systems — databases, APIs, file systems. Define tools as function schemas that the agent's LLM can call during execution.
+### What Are Workflows?
 
-### Runs
-Every agent or workflow execution creates a Run. Runs are fully logged and can be inspected in the dashboard for debugging. Runs support both synchronous and asynchronous execution modes.
+Workflows let you chain agents together with conditional logic, parallel execution, and data passing between steps. Think of them like CI/CD pipelines but for AI tasks — each step can branch, retry, or escalate based on the previous step's output.
+
+### What Are Tools?
+
+Tools give agents the ability to interact with external systems — databases, APIs, file systems. Define tools as function schemas that the agent's LLM can call during execution. doany handles the tool call loop automatically.
+
+### What Are Runs?
+
+Every agent or workflow execution creates a Run. Runs are fully logged with input, output, intermediate steps, tool calls, latency, and token usage. Inspect Runs in the doany dashboard for debugging or via the API for programmatic monitoring.
+
+## doany vs Other AI Agent Frameworks
+
+| Feature | doany | LangChain | CrewAI | AutoGen |
+|---------|-------|-----------|--------|---------|
+| **Type** | Managed API | Open-source library | Open-source framework | Open-source framework |
+| **Setup** | API key, no infra | Self-hosted, requires setup | Self-hosted, requires setup | Self-hosted, requires setup |
+| **Multi-agent workflows** | Built-in with conditional logic | Via LangGraph (separate package) | Built-in role-based | Built-in conversation patterns |
+| **Observability** | Dashboard + API logs included | LangSmith (separate product) | Limited built-in | Limited built-in |
+| **Async execution** | Native with webhooks | Requires custom implementation | Limited | Limited |
+| **Best for** | Production APIs, fast deployment | Flexible prototyping, custom chains | Role-based agent teams | Research, conversational agents |
+| **Pricing** | Free tier available, usage-based | Free (self-hosted costs) | Free (self-hosted costs) | Free (self-hosted costs) |
+
+## Frequently Asked Questions
+
+### What programming languages does doany support?
+
+doany offers official SDKs for Python and TypeScript. The REST API can be called from any language that supports HTTP requests, including Go, Ruby, Java, and Rust.
+
+### How is doany different from calling an LLM API directly?
+
+Calling an LLM API gives you a single prompt-response pair. doany adds agent orchestration on top: multi-turn tool use, workflow chaining, conditional branching, parallel execution, async runs with webhooks, and full observability — all managed via API so you don't build this infrastructure yourself.
+
+### Does doany support GPT-4, Claude, and other models?
+
+Yes. doany agents can use any supported LLM, including models from OpenAI (GPT-4o, GPT-4), Anthropic (Claude Sonnet, Opus), and others. Specify the model when creating an agent.
+
+### What does doany cost?
+
+doany offers a free tier (100 requests/minute), a Pro tier at $49/month (1,000 requests/minute), and custom Enterprise pricing. See [doany.ai/pricing](https://doany.ai/pricing) for current details.
+
+### Can I self-host doany?
+
+doany is a managed API service. There is no self-hosted option. This is by design — you get zero infrastructure management, automatic scaling, and built-in observability without running any servers.
 
 ## What's Next
 
 - Read the full [API Reference](api-reference.md)
-- Browse example agents in our GitHub repo
-- Join our Discord community for help and discussion
+- Browse example agents in our [GitHub repo](https://github.com/doany-ai)
+- Join our [Discord community](https://discord.gg/doany) for help and discussion