Build, run, and scale production AI agents with intelligent model routing.
One SDK. Any model. Automatic optimization.
Website · Docs · Dashboard · npm
Agentlify is an AI agent platform that combines intelligent model routing with multi-step agent workflows. We sit between your application and 30+ AI providers, automatically selecting the best model for each request based on cost, speed, quality, and environmental impact.
Think of us as "Netlify for LLMs" — you focus on building features, we handle the infrastructure.
const Agentlify = require('agentlify-js');
const client = new Agentlify({
apiKey: process.env.AGENTLIFY_API_KEY,
routerId: 'your-router-id',
});
// One SDK, any model — Agentlify routes to the best one
const completion = await client.chat.create({
messages: [{ role: 'user', content: 'Analyze this contract...' }],
});Automatic model selection across OpenAI, Anthropic, Google, Meta, and 100+ models. Our routing engine analyzes each request and optimizes for your priorities — cost, speed, quality, or carbon footprint.
- OpenAI-compatible API — Drop-in replacement, change one line of code
- Automatic fallbacks — Production-grade reliability with intelligent retries
- Real-time analytics — Cost tracking, latency monitoring, and optimization insights
- Carbon-aware routing — Track and minimize the environmental impact of every request
Build agents with structured workflows that break complex tasks into discrete steps:
User Input → Research → Plan → Execute → Review → Output
Each step can invoke knowledge skills on-demand for domain expertise — code review methodology, security audit checklists, TDD workflows, and more.
- 20+ ready-to-deploy agent templates — Code review, debugging, API design, content writing, data analysis
- 21 knowledge skills — Structured instruction packs covering development, security, design, analysis, planning
- Visual agent builder — Customize steps and skills without code
- Runs & observability — View, replay, and export agent execution traces
- JavaScript/TypeScript SDK — Full type safety, streaming, function calling
- MCP Server — Integrate Agentlify directly into AI coding assistants (Windsurf, Cursor, Cline)
- Local tool callbacks — Execute agent tools securely in your own code
- BYOK support — Bring your own API keys for any provider
| Repository | Description |
|---|---|
| agentlify-js | JavaScript/TypeScript SDK — OpenAI-compatible client library |
| agentlify-mcp | MCP server for AI coding assistants |
| agentlify-docs | Documentation site and API reference |
Pre-built, production-ready agent workflows you can fork and customize:
Multi-Step Agents — Professional workflows with knowledge skill integration
| Agent | Category | What It Does |
|---|---|---|
| Code Review Agent | Development | Reviews code for quality, security, and best practices |
| Bug Investigation Agent | Development | Systematically finds root causes using debugging methodology |
| Test Writer Agent | Development | Creates comprehensive test suites following TDD |
| Refactoring Agent | Development | Safely refactors code with incremental verification |
| API Designer Agent | Development | Designs RESTful APIs with full documentation |
| Security Auditor Agent | Security | OWASP-based security audits with remediation guidance |
| Documentation Writer | Content | Creates technical docs, READMEs, and API guides |
| Data Analyst Agent | Analytics | Structured data analysis with insight reports |
| Support Triage Agent | Support | Classifies and routes support tickets with draft responses |
| Content Writer Agent | Content | Creates blog posts, emails, and marketing copy |
Simple Agents — Lightweight prompt chains for focused tasks: chatbot assistant, code debugger, content summarizer, translator, email writer, SQL builder, and more.
Structured instruction packs that teach agents how to think about a problem domain:
| Category | Skills |
|---|---|
| Development | Code Review, Systematic Debugging, TDD, Refactoring, Error Handling, Performance Optimization, Code Execution |
| Design | API Design, Agent Workflow Design, Database Operations |
| Content | Content Generation, Technical Documentation, Output Formatting |
| Analysis | Data Analysis, Research Synthesis, Information Extraction |
| Planning | Writing Implementation Plans, Decision Making Under Uncertainty |
| Security | Security Review (OWASP Top 10) |
| Support | Support Ticket Triage |
| Quality | Self-Review & Revision Loop |
npm install agentlify-jsconst Agentlify = require('agentlify-js');
const client = new Agentlify({
apiKey: 'your-api-key',
routerId: 'your-router-id',
});
// Simple chat completion with automatic routing
const response = await client.chat.create({
messages: [{ role: 'user', content: 'Hello!' }],
});
// Run a multi-step agent with local tools
const agentResponse = await client.agents.run({
agentId: 'code-review-agent',
messages: [{ role: 'user', content: 'Review this PR for security issues' }],
tools: [{
type: 'function',
function: { name: 'get_pr_diff', description: 'Fetch PR diff' },
callback: async (args) => await fetchPRDiff(args.prId),
}],
});┌─────────────────┐ ┌──────────────────────────────────┐ ┌─────────────────┐
│ │ │ Agentlify │ │ AI Providers │
│ Your App │ │ │ │ │
│ │ │ ┌────────────┐ ┌────────────┐ │ │ OpenAI │
│ agentlify-js │─────┼──│ Router │──│ Agent │──┼─────│ Anthropic │
│ │ │ │ Engine │ │ Runtime │ │ │ Google │
│ │ │ └────────────┘ └────────────┘ │ │ Meta │
└─────────────────┘ │ │ │ │ │ 100+ more │
│ ┌────┴───────────────┴───────┐ │ └─────────────────┘
│ │ Analytics · Skills · Runs │ │
│ └────────────────────────────┘ │
└──────────────────────────────────┘
- Website: agentlify.co
- Dashboard: agentlify.app
- Documentation: docs.agentlify.co
- npm: agentlify-js
- Email: help@agentlify.co
Built with care in Boston