September 2026 Summaries

Vonage’s Q3 2026 developer recap highlights updates to learning resources, AI-assisted tooling, security, authentication, and video APIs. The public-beta Learning Center offers use-case examples and interactive learning paths for Video, Voice, Verify, and Messages APIs, with general availability planned for Q4, while MCP Server releases add AI-assistant access to reporting data, phone-number provisioning, application creation, chart generation, and reliability improvements. Security releases include an Okta connector that routes OTP delivery through Vonage Verify with fallback, fraud protections, and optional identity checks, alongside passkey login support based on FIDO2/WebAuthn for phishing-resistant passwordless authentication. The Video Client SDK 2.35.1 introduces Promise and async/await support in JavaScript, configurable scalable video coding layer selection, Android 17 validation, and a WebRTC M145 upgrade for native SDKs.

Phone numbers entered during e-commerce checkout can be malformed, stored in inconsistent formats, or linked to landlines and other lines unable to receive SMS, potentially causing missed delivery notifications and avoidable “Where Is My Order” support requests. The Vonage Identity Insights API is presented as a way to address these issues by using its format insight to normalize numbers to E.164 and its current_carrier insight to check live carrier-routing data and determine whether a number is currently a mobile line. Unlike regex validation and static carrier databases, which cannot reliably identify ported numbers or distinguish all line types, the API provides current carrier information globally without Network Registry approval for these insights. The article recommends invoking the API server-side at checkout before storing the number or sending messages, caching results during the session, and maintaining timeout fallbacks; it also notes JWT-based authentication through a Vonage application and suggests Vonage Verify for OTP verification of higher-risk orders.

On September 12, 2026, more than 100 participants gathered at LUMA Studios in New York City for the Runway to Reality AI Fashion Hackathon, where teams used Google Gemini AI, Antigravity, and Vonage’s Video API to create fashion technology demos in one day. Following a keynote on fashion and technology and a hands-on workshop, participants submitted 19 projects addressing virtual fitting, sustainable material use, thrift-store shopping, and personalized styling. DressMe won first place for its virtual fitting room that analyzes body measurements and color palettes from photos, while Deadstock Live Lab and ThreadLine placed second and third for sustainable surplus-fabric design and live thrift-store consultations, respectively. FitSwipe received the best API-use award for a swipe-based luxury outfit curation platform using live camera input and AI-generated looks. The event also featured an AI photobooth built with Gemini, Google Cloud services, and the Vonage Messages API, allowing attendees to generate virtual outfits, receive them through RCS or SMS/MMS, and share images on LinkedIn.

A Vonage tutorial shows how to extend a Deepgram-powered phone voice agent beyond conversation by adding a controlled order-status lookup tool, strict latency limits, deterministic failure handling, human transfer, and call logging. The Node.js example registers a client-side Deepgram function that lets the application, rather than the model, authorize and execute a mocked order lookup, allowing only one request per call and retrying only temporary transport failures. To avoid silence during live calls, lookups time out after 1.5 seconds and trigger a fixed spoken fallback, while unknown orders remain valid lookup results. Requests involving billing, returns, cancellations, or account issues are identified through predefined keywords and transferred through Vonage to a support number using an NCCO. Each interaction is saved to SQLite with transcripts, tool activity, latency, fallback and handoff details, outcomes, and agent version information, enabling teams to identify recurring problems and evaluate future capabilities.

Nimrod Taiblum of Vonage examines the often-unreported input-token costs of using JSON schemas for structured LLM output, finding that schemas can substantially increase billing for short, high-volume tasks such as classification, routing, and entity extraction. Tests using five schema complexities with Gemini 2.5 Flash, Gemini 3.5 Flash, and Claude Haiku 4.5 showed that overhead rises with schema complexity but varied sharply by model: Gemini 2.5 Flash remained near the prompt-only baseline, while Claude Haiku 4.5 reached 1,189 input tokens for a deeply nested schema versus 186 for Gemini 2.5 Flash. For a roughly 100-token prompt with a medium schema, reported costs per million calls ranged from $37.20 for Gemini 2.5 Flash to $396 for Gemini 3.5 Flash, reflecting both token prices and differing schema handling. The analysis suggests that schema overhead becomes relatively insignificant for long inputs but can dominate costs for prompts under 500 tokens, and recommends minimizing and flattening schemas, placing explanatory details in prompts where appropriate, batching requests, considering schema-free JSON prompting for trivial outputs, and measuring schema-related token differences when selecting models and estimating production costs.

Vonage has added three healthcare-focused AI companies to its Startup Program, highlighting their use of voice, messaging, and verification APIs to improve patient access and administrative workflows. MedFeed AI is deploying multilingual phone and WhatsApp assistants in Indian hospitals to automate patient feedback, outcome measurement, and accreditation compliance, reporting improved satisfaction results at two hospitals. SurgePoint provides aesthetics practices with an AI front desk that manages HIPAA-compliant calls, appointment inquiries, SMS conversations, and rebooking campaigns. Founded by a practicing physician, Switchboard, MD consolidates referrals from fragmented channels such as faxes, voicemails, emails, and online forms, then uses AI voice and SMS outreach to help specialty providers convert referrals into appointments. Vonage positions the program as support for early-stage teams using communications technology to address healthcare and other industry challenges.

Vonage’s China Relay service, which provided mainland China-based geo-detection, signaling proxying, and TURN media relaying through AWS Beijing, was permanently retired in Q2 2026. Customers serving users in China are advised to create and operate their own compliant relay infrastructure because cross-border WebRTC signaling and media can face filtering, latency, and packet loss, while Chinese internet-facing services require appropriate ICP authorization. The proposed approach combines a customer-hosted TURN server and HTTP/WebSocket reverse proxy on a Chinese cloud provider with Vonage Video API’s Configurable TURN and IP Proxy add-ons, which respectively route media and non-media signaling traffic through those local systems. Organizations must obtain required licensing, deploy and secure infrastructure with stable Chinese IP addresses and domains, configure client SDKs to force custom TURN and use the proxy, and validate operation across Chinese ISPs, platforms, server logs, and WebRTC relay statistics.

Vonage’s Verify RCS Default Agent is a centrally managed, pre-registered RCS sender that lets eligible Verify V2 API customers in supported countries send RCS authentication messages immediately without waiting for carrier approval or completing custom agent registration. Developers enable it by omitting the `from` parameter, setting the workflow channel to RCS, and using the standard Verify API process to initiate and validate one-time passcodes with a returned request ID. Messages are sent from the verified “Vonage Verify” agent, while the customer’s brand can appear in the message body; organizations can register their own branded agent in parallel for a later transition. The service supports fallback workflows that attempt RCS first and SMS second, with Vonage handling channel selection automatically. Compared with SMS, RCS can offer rich OTP cards, copy-code controls, autofill support on compatible devices, and a carrier-verified sender identity, while pricing, general API usage limits, and billing remain the same as standard Verify RCS.

Marking PHP’s 31st year, a Vonage developer advocate summarizes the first half of JetBrains’ PHPVerse 2026 conference, presenting PHP as a mature ecosystem focused on governance, security, modernization, and sustainable open-source maintenance. PHP Foundation executive director Elizabeth Barron described community feedback identifying transparency, public perception, developer onboarding, and internal voting as priorities, leading to special-interest groups for security, outreach, cryptography, accessibility, events, and onboarding. Composer and Packagist co-creator Nils Adermann examined supply-chain risks and outlined existing and planned protections, including immutable releases, dependency policies, scanning, transparency logs, MFA, staged publishing, and artifact hosting, while stressing Packagist’s large-scale infrastructure and need for sponsorship. Jonathan Bossinger argued that modern WordPress has addressed common criticisms through PHP 8.3 support, modern object-oriented features, block-based templating and state management, performance improvements, security measures, Composer-oriented tools, upgraded documentation, and native exception support. Symfony creator Fabien Potencier detailed how the framework manages hundreds of repositories and thousands of releases through a stable core team, opinionated conventions, semantic versioning, predictable release cycles, repository splitting, and AI-assisted automation, while noting that AI-driven vulnerability discovery has increased both the pace of security reporting and the need to validate findings.

Vonage’s tutorial explains how to reduce failed WhatsApp deliveries by validating phone numbers with the Identity Insights API before sending messages. It recommends requesting both the format insight, which confirms structural validity and normalizes numbers to E.164 format, and the current carrier insight, which identifies the carrier and network type; only numbers with a valid format and a MOBILE network type should receive WhatsApp messages. Developers can explore these checks in the Identity Insights Playground or call the API through cURL using JWT authentication. The tutorial also describes configuring the Messages API Sandbox, including allowlisting a WhatsApp number and sending a test message through its sandbox endpoint. Finally, it shows how the Vonage Tooling MCP Server can connect these steps to AI clients such as Claude or Gemini, enabling a natural-language prompt to validate a number and send a WhatsApp message only when the validation succeeds.

A Vonage tutorial presents a layered authentication approach for reducing account takeover fraud in FastAPI and Kotlin applications by combining SIM swap detection, silent network authentication, and OTP fallbacks. It explains that SMS OTP alone is vulnerable to SS7 interception, phishing, and SIM swap attacks, while carrier-side silent authentication verifies a device’s active SIM through a cellular-data request without requiring user input. Using Vonage Identity Insights, the sample first checks whether a phone number’s SIM was recently changed; suspicious or unavailable results trigger an email OTP, while numbers without a recent swap proceed to Vonage Verify silent authentication. If silent authentication is unsupported or fails, such as when the device is on Wi-Fi or the carrier lacks coverage, the application falls back to SMS OTP. The provided FastAPI backend uses the Vonage Python SDK to initiate and validate these flows, and the Kotlin Android client uses the Vonage SDK to force a cellular GET request to the supplied verification URL, returning a code for server-side validation.

Vonage outlines several support and learning channels for developers building with its APIs, recommending the Developer Portal and Learning Center for API references, code examples, SDK guidance, authentication details, and structured courses. Kapa AI, available through the support and developer portals, is positioned as a quick resource for common errors, configuration questions, and known fixes, while support tickets are intended for issues requiring tracked resolution, including production outages, unresolved API failures, billing concerns, account problems, and number provisioning. For in-person engagement, the Vonage Community Partners program offers developer meetups, workshops, hackathons, and networking opportunities searchable by location and focus area. The post advises developers with urgent production-impacting problems to submit a ticket directly and encourages ongoing engagement through newsletters, social channels, tutorials, feedback programs, and community events.