Build an AI agent on CEO.ai. Train it with RAG knowledge. Get its API key. Call it from any application, frontend, service, or system — anywhere in the world.
Your agents aren't locked inside our platform. They're infrastructure you build on. Embed intelligence into your website, your app, your Slack workspace, your client projects, or your own AI-powered products. One API call. RAG-powered responses. Your agent, your knowledge, your rules.
A URL you can call from any application, anywhere
Authentication scoped to that specific agent
The agent uses all its trained knowledge via API, exactly as in the app
That means: You build an agent in the CEO.ai app. You train it on your company's product documentation using RAG. You grab its API key. You call it from your website's chat widget.
A customer asks "What's included in the Enterprise plan?" — and your agent answers using YOUR actual pricing page, YOUR feature comparison, YOUR terms. Not a generic response. YOUR response.
The agent doesn't care where the request comes from. It responds the same way whether called from the CEO.ai app, your React frontend, a Node.js backend, a Python script, a mobile app, or a serverless function.
$ npm install @ceo-ai/sdk
const { CeoAI } = require('@ceo-ai/sdk');
const ceo = new CeoAI({ apiKey: process.env.CEO_API_KEY });
// Send a prompt and wait for the result
const { response, metadata } = await ceo.promptAndWait(
'What is our refund policy for enterprise customers?'
);
console.log(response);
// => Your exact refund policy, pulled from RAG-trained knowledgeconst { CeoAI } = require('@ceo-ai/sdk');
const ceo = new CeoAI({ apiKey: process.env.SALES_AGENT_KEY });
// First message
const first = await ceo.promptAndWait(
'I run a 30-person marketing agency and want to automate client reporting.'
);
// Follow-up with conversation history
const second = await ceo.promptAndWait(
'We use HubSpot and Google Analytics. Can you help with both?',
{
conversationHistory: [
{ role: 'user', content: 'I run a 30-person marketing agency...' },
{ role: 'assistant', content: JSON.stringify(first.response) }
]
}
);
// Each message has full context of previous messages
// + the agent's RAG knowledge about your products and qualification criteriaconst { CeoAI } = require('@ceo-ai/sdk');
const ceo = new CeoAI({ apiKey: process.env.CEO_API_KEY });
// Submit prompt — returns immediately with presigned URL
const submission = await ceo.prompt('Generate a detailed quarterly report');
console.log('Presigned URL:', submission.presignedUrl);
// Store the URL in your database...
// Later (even from a different process), retrieve the result:
const result = await ceo.pollForResult(storedPresignedUrl);
console.log(result);import requests
response = requests.post(
'https://ingest.api.ceo.ai/prompt',
headers={
'x-api-key': 'your-agent-api-key',
'Content-Type': 'application/json'
},
json={
'prompt': 'Summarize the Q3 pipeline by region'
}
)
result = response.json()
print(result)curl -X POST https://ingest.api.ceo.ai/prompt \
-H "x-api-key: your-agent-api-key" \
-H "Content-Type: application/json" \
-d '{"prompt": "What are the top 3 objections for mid-market prospects?"}'The bottom line: If your application can make an HTTP request, it can talk to your CEO.ai agents. Any language. Any framework. Any platform.
The Agent API turns every agent you create into embeddable, callable intelligence. Here's what people are building:
Build a support agent. Train it on your product docs, FAQs, and SOPs. Embed it in a chat widget on your website. Customers get instant, accurate, RAG-powered answers — without a human in the loop for the 80% of questions that are routine.
Build an agent. Train it on whatever knowledge it needs. Connect it to a messaging platform via their bot API. Now you have an intelligent bot that responds with real business knowledge — not canned responses.
Example: A Telegram bot for a real estate agency that answers property questions using the agent's RAG knowledge of current listings, pricing, neighborhood data, and availability.
Already have an internal dashboard or tool? Add an AI layer without rebuilding anything. Your existing app calls the Agent API for analysis, summarization, data transformation, or recommendations.
Example: An internal sales dashboard with an "Ask the Agent" button. The sales team types a question about a deal, competitor, or product — and gets an instant, RAG-powered answer.
Building a product? Add AI features without building ML infrastructure. Create agents that power specific features — document analysis, content generation, data extraction, intelligent search — and call them via API.
Example: A project management SaaS adds an "AI Brief Generator" feature. Behind the scenes, it calls a CEO.ai agent trained on PM best practices and the user's project history.
The Agent API lets you create, train, and sell AI agents as a business. Build specialized agents for an industry vertical. Train them on domain expertise. Sell API access — or build lightweight frontends and sell them as products.
Example: An agency builds a "Legal Document Analyzer" agent trained on contract law patterns and their firm's methodology. They sell access to other law firms via a simple web interface — powered by the Agent API on the backend.
Not everything needs a user interface. Call agents from serverless functions, cron jobs, webhooks, or backend services for automated processing.
Example: New support ticket via webhook → backend calls triage agent → agent categorizes, assesses urgency, suggests response, routes to right team — all before a human sees it.
Each agent on CEO.ai is a standalone, independently callable AI entity. This is a deliberate architecture decision:
One agent = one API key = one endpoint = one purpose
Your website might use a support agent for chat, a sales agent for pricing, and a product agent for comparisons. Each call goes to the right specialist.
Give a client the API key for their agent — they can call it without accessing any of your other agents or data.
If a specific integration is compromised, rotate that agent's key without affecting any other agent.
See which agents are getting called, how often, and from where. Understand which generate the most value.
This 1-to-1 architecture is what makes the Agent API viable for building products, services, and businesses — not just internal tools. Each agent is a self-contained AI microservice that you can deploy, manage, and monetize independently.
On Enterprise plans, you get API access to the CEO Agent itself.
This means you can programmatically:
The CEO Agent API is how you integrate full project generation into your own systems and workflows. Agencies can build client projects programmatically. DevOps teams can auto-generate infrastructure. Product teams can spin up prototypes on demand.
Available on Enterprise — Talk to Our Team$ npm install @ceo-ai/sdk
Zero external dependencies. Built on Node.js built-in modules. Prompt agents, manage conversations, train RAG memory, and poll for results. Async/await patterns. TypeScript-friendly.
ceo.promptAndWait() — send & wait for responseceo.prompt() — fire-and-forget, poll laterceo.addRag() — programmatic RAG trainingCeoAPIError — typed error handling$ npm install -g @ceo-ai/cli
Manage agents, RAG knowledge, and conversations from your terminal. Ideal for scripting, CI/CD integration, and bulk operations.
# Chat with an agent
ceo prompt "What are our Q3 numbers?"
# Add RAG knowledge — single file
ceo addRag ./docs/pricing-guide.pdf
# Add entire folder recursively
ceo addRagDirectory ./documentation/| Resource | Link |
|---|---|
| API Reference | docs.ceo.ai |
| SDK Documentation | SDK Reference |
| CLI Documentation | CLI Reference |
| npm: @ceo-ai/sdk | npmjs.com |
| npm: @ceo-ai/cli | npmjs.com |
| Capability | Startup $297/mo |
SMB $1,499/mo |
Enterprise $5,500+/mo |
|---|---|---|---|
| Agent API (per-agent API keys) | ✅ | ✅ | ✅ |
| SDK & CLI | ✅ | ✅ | ✅ |
| RAG-powered API responses | ✅ | ✅ | ✅ |
| Multi-turn conversations via API | ✅ | ✅ | ✅ |
| Workflows API | — | ✅ | ✅ |
| Webhook triggers | — | ✅ | ✅ |
| CEO Agent API | — | — | ✅ |
| API integration support (our team) | — | — | ✅ |
| Custom rate limits | — | — | ✅ |
| Credits | 10,000/mo | 50,000/mo | Custom |
Every plan includes Agent API access. If you're a developer who wants to build AI into products, Startup at $297/month gets you agent creation, RAG training, and API access with 10,000 credits. That's enough to build and test multiple AI-powered integrations.
The Agent API turns CEO.ai from a platform you use into AI infrastructure you build on. Create an agent, train it on your knowledge, and call it from anywhere — your website, your app, your product, your clients' systems.
Every plan includes Agent API access. Get started today.
Agent API included on all plans. SDK and CLI available on npm. CEO Agent API available on Enterprise.