Trending AI tools worth your attention.
A running shortlist of the AI products we have reviewed most recently, grouped by who they are actually for. Every tool here has a full write-up covering pricing, real limitations and alternatives — not a feature list copied from the vendor.
How tools get here: we add a tool when it ships something genuinely new, when readers ask about it, or when its pricing changes enough to matter. This page rebuilds itself as reviews are published, so the newest work is always at the top of each section.
For creators
Voice, video, image and design tools for people producing content.
- PetIntoArt Review 2026: Six Free Portraits, 109 Looks, and What Each One Actually CostsThree free generations with no account, three more with a free sign-up, then one-time credit packs from $9.99. We counted the 109 Looks and worked out the cost per portrait.
- ChatGPT Images 2.5 Review 2026: Sketch, Faster Generation, Precision Editing and API ModelsChatGPT Images 2.5 launched 8 September 2026 with Sketch, templates, image comments and prompt sharing. Its two API models, Flare and Sunburst, are priced identically — so the choice is latency against detail.
- ElevenLabs Quietly Made v4 Turbo the Default for Every New Voice AgentOn 5 October 2026 ElevenLabs changed the default conversational model to eleven_v4_turbo and added gpt-6.1-sol and claude-sonnet-5-5 as agent LLMs. It went out as a changelog line, not an announcement.
- Fireflies Just Raised Its API Limits 10x — and the Free Plan Is Now Genuinely UsableFireflies raised API rate limits tenfold on 8 October 2026: free goes 50 to 500 requests a day, Pro 500 to 5,000. Business and Enterprise moved to pooled per-seat limits with a ceiling.
- BytePlus September 2026: Seedream 5.0 Flash, Seedance 2.5 Premium and Two New Reasoning ModelsBytePlus added Seedream 5.0 Flash and two reasoning models to ModelArk in September, and shipped Seedance 2.5 Premium plus layer decomposition in its official ComfyUI plugin on 29 September. Model IDs, what each one does, and what is still unverified.
- Fireflies Talk vs Wispr Flow (2026): Free Unlimited Dictation vs 2,000 Words a WeekFireflies Talk launched on 29 September 2026 with free, unlimited dictation on every plan. Wispr Flow caps free users at 2,000 words a week. We checked both pricing pages and compared what you actually get.
For builders and founders
No-code app builders, AI development platforms and automation.
- Revamp Review 2026: Redesign From a URL, and the Domain Cap That Decides the PriceRevamp rebuilds your existing site from its URL in a preview while the old one stays live. Studio cannot publish to your own domain, and Grow includes only three — then it is $15 a month each.
- 21st.dev Shipped Video in Three Days: Remotion Blocks, a Codex Plugin, Then the Fixes21st.dev added Remotion video blocks on 7 October, a Codex plugin that turns a prompt into an editable video on 8 October, and fixed the playback on 9 October.
- n8n 2.43: Browser Use Opens to Everyone, and the Supabase Line Matters More Than It Looksn8n 2.43 took Browser Use off the gate for all users, with automatic credential setup off by default. Plus Supabase OAuth2 that also enables the Supabase MCP server.
- Bubble Agent 2 Leaves Beta, the MCP Is Unmetered, and Starter Goes Up on 20 OctoberBubble shipped Agent 2, an unmetered MCP server, AI credits on Agent edits and a Starter price rise in one 6 October announcement. The deadline to lock $29/mo is 19 October.
- Hostinger Adds Google Ads With $500 Credit, WhatsApp Campaigns and a Three-Tool MCPHostinger’s Week 40 update adds a Google Ads integration carrying $500 in Google ad credit after $500 of spend, WhatsApp marketing campaigns in Reach, and an MCP server collapsed to three tools. Verified prices and the terms nobody reads.
- Replit’s October 2026 Update: GPT-6.1 Sol, Claude Sonnet 5.5 and the Jev Decision ModelReplit’s 2 October 2026 changelog added GPT-6.1 Sol in Max Mode, Claude Sonnet 5.5 in Power Mode, the Jev decision model via AI Integrations, and a full-page Settings. Verified prices, mode gating and the OpenRouter privacy caveat.
For marketers
Writing, campaign and content tools for small teams.
- How Kerbrank Measures Local Visibility — and the One Metric That Flatters Every BusinessKerbrank publishes its full scanning method. Read it and you learn how to judge any local rank tracker — starting with ARP, the metric that scores 1.0 for a business visible at 2 points out of 49.
- Kerbrank Review 2026: UK Local Rank Tracking, ChatGPT Checks, and the Cap Nobody ReadsKerbrank tracks Google Maps visibility across a grid and samples ChatGPT recommendations, from £29 a month with no scan credits. The limit that decides what you can measure is map checks a week — and it peaks at Pro.
- Hostinger Adds Google Ads With $500 Credit, WhatsApp Campaigns and a Three-Tool MCPHostinger’s Week 40 update adds a Google Ads integration carrying $500 in Google ad credit after $500 of spend, WhatsApp marketing campaigns in Reach, and an MCP server collapsed to three tools. Verified prices and the terms nobody reads.
Assistants and research
General-purpose AI assistants and research tools.
- ChatGPT Pricing in October 2026: the New $500 Pro Tier, the Pro 200 Reset and Every Published LimitOpenAI launched a $500 ChatGPT Pro tier on 29 September 2026 and reopened Pro 200 with smaller allowances. Every tier price, every published message limit, the Chat vs Work model split, and the custom GPT deadlines.
- Typeform’s MCP Server: Build, Theme and Publish Forms From Claude or ChatGPTTypeform’s September 2026 update lets Claude list and apply form themes through the Typeform MCP server. Here is what the server can do, the three server URLs, and the personal-access-token change that breaks existing setups.
- ChatLLM by Abacus.AI Review (2026): 100+ Models for $10 a Month — and the Credit System Nobody ExplainsChatLLM by Abacus.AI bundles 100+ AI models, an agent, a coding IDE and image and video generation for $10 per user per month. We checked the pricing page, the credit limits and the complaints.
- Claude Opus 5.5 Is Not 40% Cheaper — What the List Price Actually SaysAnthropic’s Opus 5.5 headline says 40% cheaper. List prices fell 20%, cache reads fell 60%, and the rest depends on your workload. The numbers, with worked examples.
- Ads Are Coming to ChatGPT: Which Plans Show Them and What It Costs to Opt OutAI Assistant guide Ads Are Coming to ChatGPT: Which Plans Show Them and What It Costs to Opt Out We have no… Read more: Ads Are Coming to ChatGPT: Which Plans Show Them and What It Costs to Opt Out
- Custom GPTs Are Being Retired: Dates, What Transfers, and What You RebuildAI Assistant guide Custom GPTs Are Being Retired: Dates, What Transfers, and What You Rebuild Deadline: you can no longer create new… Read more: Custom GPTs Are Being Retired: Dates, What Transfers, and What You Rebuild