Developers

An API built for production agents.

Everything in the dashboard is available programmatically. Create agents, stream chats, index knowledge, and read usage from your own systems.

API keys

Scoped live and test keys per workspace, rotatable without downtime.

Rate limits

Standard limit headers on every response, with dedicated capacity on Enterprise.

Webhooks

Signed event delivery for conversations, escalations, and usage thresholds.

SDKs

Typed clients for Python, Node.js, and the browser, generated from OpenAPI.

API Platform

Nine surfaces for building on AgentLabz

The console is a client of the same API you get. Anything you can configure in the product, you can automate.

REST API

Resource-oriented endpoints for agents, runs, knowledge, and voice sessions.

Python SDK

Typed client with async support, retries, and streaming helpers.

JavaScript SDK

Isomorphic TypeScript client for browser, Node, and edge runtimes.

Streaming API

Server-sent events for tokens, tool calls, and run state transitions.

Webhooks

Signed delivery with retries, replay, and per-event subscriptions.

OpenAPI Docs

A published spec that generates clients and stays in sync with the API.

Rate Limits

Standard headers, per-key quotas, burst allowances, and dedicated capacity.

API Keys

Scoped keys per environment with rotation and last-used visibility.

Playground

Run an agent against real inputs in the browser and copy the request.

Core endpoints

POST/v1/agentsCreate an AI employee with instructions, tools, and model settings.
GET/v1/agents/{id}Retrieve an agent's full configuration.
POST/v1/chatsStart a conversation with an agent and stream the response.
POST/v1/knowledge/documentsUpload a document for chunking, embedding, and indexing.
GET/v1/analytics/usagePull conversation, token, and cost usage for a period.
POST/v1/webhooksSubscribe to conversation, escalation, and billing events.

SDK examples

Python

from agentlabz import AgentLabz

client = AgentLabz(api_key="al_live_...")

reply = client.chats.create(
    agent_id="agent_support",
    message="Where is my order #48219?",
    stream=True,
)

for chunk in reply:
    print(chunk.text, end="")

Node.js

import { AgentLabz } from "@agentlabz/sdk";

const client = new AgentLabz({ apiKey: process.env.AGENTLABZ_API_KEY });

const stream = await client.chats.create({
  agentId: "agent_support",
  message: "Where is my order #48219?",
  stream: true,
});

for await (const chunk of stream) {
  process.stdout.write(chunk.text);
}

cURL

curl https://api.agentlabz.tech/v1/chats \
  -H "Authorization: Bearer $AGENTLABZ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "agent_id": "agent_support",
    "message": "Where is my order #48219?"
  }'
Documentation

Everything you need to ship

Guides, references, and runnable examples for every part of the platform.

app.agentlabz.tech / settings / api keys

production

alz_live_••••••••••••7f2a

used 2 min ago

staging

alz_test_••••••••••••91c4

used 1 day ago

ci-pipeline

alz_test_••••••••••••4b08

used 3 days ago
Rate limit600 req / min
Rotation policy90 days
Scopesagents:run · knowledge:read

Quick Start

Deploy your first agent and send a message in under ten minutes.

Authentication

API keys, scopes, rotation, and workspace-level access.

Deploy Agent

Configure, evaluate, promote, and roll back an agent version.

Knowledge Base

Ingestion, chunking, re-indexing, and permission-aware retrieval.

Voice API

Sessions, telephony routing, transcripts, and latency tuning.

Workflows

Triggers, branching, retries, and durable state machines.

SDKs

Python and JavaScript clients with streaming and typed models.

CLI

Manage agents, secrets, and deployments from your terminal.

Examples

Reference implementations for support, sales, voice, and RAG.