Explore the latest company news, product updates, technical tutorials, and best practices from Twilio, the industry-leading customer engagement platform.
Step-by-step tutorial showing how to implement passwordless magic-link login in Laravel using Twilio SMS. Covers project setup, database migration and Eloquent model, configuring Twilio credentials and installing its PHP SDK, building controller logic for registration and magic-link authentication, creating Blade views and routes, running the app with ngrok, and testing the SMS-based login flow.
A technical guide to core latency in cascaded AI voice agents (STT → LLM → TTS). It defines mouth-to-ear vs platform latency, provides a typical latency breakdown and initial p50/p95 targets, and gives engineering recommendations—colocate services at media edges, avoid unnecessary network traversals, prefer streaming LLM APIs and benchmark TTFT, tune end-of-turn detection with graceful aborts, and weigh TTS expressiveness against raw latency. The post also advocates managed orchestration (Twilio ConversationRelay) for low-latency voice AI deployments.
Step-by-step tutorial showing how to integrate Twilio SMS with Zendesk: buy a Twilio number, create a Zendesk webhook and trigger to send outbound SMS via Twilio, and build a Twilio Function (using Zendesk API token and environment variables) to route incoming SMS into Zendesk tickets. Includes deployment and configuration notes plus suggested next steps (WhatsApp, Voice, Twilio Studio).
A hands‑on CakePHP tutorial showing how to build a health-focused chatbot that uses Twilio's WhatsApp API for messaging and Google Gemini for AI responses. It covers project setup, environment variables, installing Twilio's PHP helper, creating a TwilioController to handle webhooks and call Gemini, configuring the Twilio Sandbox and ngrok, disabling CSRF for the webhook, and testing the bot; it also demonstrates a strict health-focused prompt to constrain AI replies.
Twilio automated generation and maintenance of accessible alt text and consistent image filenames for their documentation by combining LLM-driven prompt engineering with tooling: a CLI batch processor, a Docs AI Buddy review web app, and a remark plugin to update markdown. The system processed 2,734 images with 99.96% human-approval on generated alt text, saving hundreds of hours and improving accessibility and consistency.
