Getting Started
Build and run your first AI agent in 5 minutes.
Prerequisites
- Sekuire CLI installed
- API key from your dashboard
- LLM API key (OpenAI, Anthropic, etc.)
Step 1: Install the CLI
macOS / Linux:
curl -fsSL https://install.sekuire.ai | sh
Verify installation:
sekuire --version
Step 2: Authenticate
sekuire login
This prompts you for an API key. Get one from your dashboard at /dashboard/settings?tab=api-keys.
Step 3: Create a Project
sekuire init --name my-agent --language typescript
cd my-agent
This creates:
sekuire.yml- Agent configurationsystem_prompt.md- LLM system prompt.sekuire/- Keypair (never commitsecret.key)
Step 4: Configure Your Agent
Edit sekuire.yml:
sekuire.yml
project:
name: "my-agent"
version: "1.0.0"
agents:
assistant:
name: "AI Assistant"
system_prompt: "./system_prompt.md"
llm:
provider: "openai"
model: "gpt-4-turbo"
api_key_env: "OPENAI_API_KEY"
temperature: 0.7
Set your LLM API key:
export OPENAI_API_KEY="sk-..."
Step 5: Install SDK & Run
- TypeScript
- Python
- Rust
npm install @sekuire/sdk
// main.ts
import { getAgent } from '@sekuire/sdk';
const agent = await getAgent('assistant');
const response = await agent.chat('Hello!');
console.log(response);
pip install sekuire-sdk
# main.py
import asyncio
from sekuire_sdk import get_agent
async def main():
agent = await get_agent('assistant')
response = await agent.chat('Hello!')
print(response)
asyncio.run(main())
# Cargo.toml
[dependencies]
sekuire-sdk = "0.1"
tokio = { version = "1", features = ["full"] }
// main.rs
use sekuire_sdk::get_agent;
#[tokio::main]
async fn main() -> anyhow::Result<()> {
let mut agent = get_agent(Some("assistant"), None).await?;
let response = agent.chat("Hello!", None).await?;
println!("{}", response);
Ok(())
}
Step 6: Push to Registry
sekuire push
This publishes your agent as a draft. You can verify it with:
sekuire verify
tip
Use sekuire run to run your agent with automatic bootstrap, heartbeat, and workspace registration.
Next Steps
- Configuration - sekuire.yml deep dive
- Tools - Add capabilities
- Deployment - Production options