Google doesn’t usually swap a core image model inside Gemini without a clear reason.

Nano Banana 2 brings noticeably faster generation into the stack — and speed is one of the few upgrades that actually changes how often creators experiment and publish.

At the same time, Figma is wiring in OpenAI’s Codex, Read AI is launching an email-based digital twin, and AI guidance is showing up in more consumer apps.

Today’s updates aren’t about flashy capabilities. They’re about reducing the pause between idea and execution.

Google has released Nano Banana 2, technically Gemini 3.1 Flash Image, a new image model that keeps Pro-level fidelity while generating visuals faster at resolutions from 512px up to 4K in multiple aspect ratios. It’s now the default image engine across Gemini and Google’s Flow video editor, powers visual results in Search and Lens in 141 countries, and ships with SynthID watermarks and C2PA Content Credentials support for provenance.​

Meeting-notes startup Read AI has launched Ada, an email-based “digital twin” that can find meeting times, reply in threads with your availability, and draft responses using your calendar, past meetings, company knowledge base, and web search. Users set it up by emailing [email protected], and the assistant will soon expand beyond email to Slack and Teams as the company pushes past 5 million monthly active users.​

Bumble is rolling out AI-driven profile guidance that gives personalized feedback on your bio and prompts globally, plus a U.S.-only photo feedback tool that suggests more flattering, varied images and flags shots like sunglasses selfies that hide your face. The app is also testing a non-AI “Suggest a Date” nudge in Canada to help matches signal they’re ready to meet IRL when chats stall.​

Figma is integrating OpenAI’s Codex so teams can move fluidly between design files and code, using Figma’s MCP server to share context across both tools. Engineers can visually iterate on interfaces without leaving their coding environment, while designers work closer to real implementation, extending Figma’s recent push into AI workflows alongside its Anthropic Claude Code partnership.​

Jeetro.com has launched an AI tools discovery and indexing platform that organizes thousands of AI products across categories like writing, image and video generation, marketing, analytics, coding, and productivity. The directory lets users search, filter, and browse AI software by function and use case, aiming to cut through market fragmentation and give startups, enterprises, and creators a systematic way to evaluate tools—including popular ChatGPT alternatives.

Werd is an AI content strategy and writing platform built on a multi-agent architecture designed to plan and execute content like a full-stack team.

You describe the goal → the agents handle research, SEO planning, and drafting.

The platform leans heavily into what it calls “vibe writing,” where you guide tone and intent conversationally rather than through complex prompts.

Werd fits best if content is part of your growth engine.

Strong fit:

Less ideal:

Where Werd actually stands out:

Because multiple agents collaborate behind the scenes, the tool aims to compress the ideation → research → draft loop into one flow.

Werd is less interesting as a writer and more interesting as a planner. If your bottleneck is staring at a blank page, many tools can help. But if your bottleneck is figuring out what to publish consistently, Werd’s multi-agent approach starts to earn its keep.

Feed it real positioning context: Werd performs best when your business profile is fully filled out — the agents rely heavily on that context.

ReelFarm is an AI TikTok automation tool that generates and auto-posts slideshow-style TikToks at scale to drive traffic back to your site, complete with scheduling, analytics, Pinterest image sourcing, and AI UGC selfies for product-style content.​

Gling is an AI video editor for YouTubers that auto-cuts bad takes, silences, and filler words, then adds captions, smart zooms, and noise removal so you can export polished videos or timelines into Premiere, Final Cut, and DaVinci in a fraction of the time.​

Textideo is an AI video generator for bloggers and influencers that turns text prompts or images into short, platform-ready videos using models like Google’s Veo 3, with built-in text-to-image, image-to-video, voiceovers, and effects for TikTok, Reels, and YouTube.​

Pooks.ai

Pooks.ai creates personalized non-fiction “pooks” — full-length ebooks tailored to your age, goals, preferences, and context (from fitness to self-help and cookbooks), so you get a custom book instead of reading one-size-fits-all advice.​

AskCory is an AI marketing assistant that researches your market and then generates tailored strategies, channel tactics, KPIs, benchmark-backed plans, ready-made content, and even PowerPoint decks using the Bullseye Framework, aimed at startups, marketers, and agencies.