# Nowrap

> Discover AI Tools, Track AI News — Discover AI tools that actually do the job and track AI news that matters — curated by profession, vetted for trust, updated weekly.

Nowrap is an editorial directory of AI tools and AI news for working professionals (teachers, lawyers, doctors, marketers, designers, writers, researchers, engineers, freelancers, and real estate). Tool listings are verified by Nowrap editors for an active website, a published privacy policy, accurate pricing, and ongoing maintenance — we do not publish hands-on reviews or star ratings.

## Tools
- [Abridge](https://nowrap.ai/tools/abridge): Enterprise AI for clinical conversations.
- [Amazon Quick](https://nowrap.ai/tools/amazon-quick): An AI assistant for workplace workflows.
- [Claude Projects](https://nowrap.ai/tools/claude-projects): A long-context workspace for your work.
- [Consensus](https://nowrap.ai/tools/consensus): AI-powered search for academic papers.
- [Cursor](https://nowrap.ai/tools/cursor): An AI-first IDE.
- [Descript](https://nowrap.ai/tools/descript): Edit audio and video like a doc.
- [Diffit](https://nowrap.ai/tools/diffit): Differentiated reading materials for any classroom.
- [DoxGPT](https://nowrap.ai/tools/doxgpt): Free AI for verified US clinicians.
- [Figma Make](https://nowrap.ai/tools/figma-make): Prompt to working prototype.
- [Fireflies](https://nowrap.ai/tools/fireflies): An AI notetaker for your meetings.
- [Harvey](https://nowrap.ai/tools/harvey): AI for legal and professional services teams.
- [Jasper](https://nowrap.ai/tools/jasper): An execution platform for marketing teams.
- [Lavender](https://nowrap.ai/tools/lavender): An email coach for sales reps.
- [MagicSchool](https://nowrap.ai/tools/magicschool): 80+ AI tools, made for teachers.
- [Perplexity](https://nowrap.ai/tools/perplexity): An answer engine with citations.
- [Spellbook](https://nowrap.ai/tools/spellbook): Contract drafting and review, inside Word.
- [Workspace Agents in ChatGPT](https://nowrap.ai/tools/workspace-agents-in-chatgpt): Shared AI agents for team workflows.

## News
- [OpenAI previews GPT-5.6 Sol, Terra, and Luna with stronger coding and cyber benchmarks](https://nowrap.ai/news/openai-gpt-5-6-preview): GPT-5.6 is real, but not broadly available yet: OpenAI is starting with a limited partner preview before expanding access to ChatGPT, Codex, and the API.
- [Netflix accidentally shipped its CLAUDE.md instructions](https://nowrap.ai/news/netflix-claude-md-leak): Screenshots from the Netflix iOS app bundle appear to show a bundled Claude Code instruction file. The leak is less about secrets and more about how major companies are starting to steer AI coding agents.
- [OpenAI and Broadcom unveil Jalapeño, a custom chip built for LLM inference](https://nowrap.ai/news/openai-broadcom-jalapeno-inference-chip): OpenAI's first silicon — co-designed with Broadcom in nine months — targets 50% lower inference cost than Nvidia GPUs and signals a full-stack ambition that runs from models to data-center hardware.
- [Anthropic launches Claude Tag, an AI teammate that lives in your Slack](https://nowrap.ai/news/claude-tag-slack-ai-teammate): Claude Tag turns @Claude into a shared, always-on AI teammate inside Slack — one that learns your team's workflows, monitors channels, and picks up tasks without being handed off repeatedly.
- [SpaceX moves to buy Cursor in $60B stock deal](https://nowrap.ai/news/spacex-cursor-acquisition): SpaceX's planned acquisition of Anysphere would put Cursor's coding-agent surface, enterprise developer base, and model-training roadmap inside the SpaceXAI stack.
- [Claude Fable 5's system prompt leaked. Here's what we could verify.](https://nowrap.ai/news/claude-fable-5-system-prompt-leak): A 122,750-character file claiming to be Fable 5's system prompt is circulating on GitHub. We downloaded it and checked the claims line by line — separating what holds up from what doesn't.
- [Anthropic pulls Fable 5 after US order](https://nowrap.ai/news/anthropic-fable-5-mythos-5-shutdown): The company says a U.S. export-control directive forced it to disable Claude Fable 5 and Mythos 5 only days after launch.
- [OpenAI adds reset banking for Codex users](https://nowrap.ai/news/openai-codex-rate-limit-reset-banking): Eligible ChatGPT Plus and Pro users can now bank Codex rate-limit resets, including one free reset at launch and referral-based resets during a limited June promotion.
- [Anthropic releases Claude Fable 5 and Mythos 5](https://nowrap.ai/news/claude-fable-5-mythos-5): The first public Mythos-class models — more capable than Opus on long, complex work — ship with hard limits that fall back to Opus 4.8 on risky topics.
- [OpenAI Codex Sites turns prompts into hosted apps](https://nowrap.ai/news/openai-codex-sites-hosted-apps): Sites is a Codex preview that lets Business and Enterprise teams create, save, deploy, and inspect hosted websites and internal tools from Codex.
- [ASUS just turned ProArt into an RTX Spark AI workstation](https://nowrap.ai/news/asus-proart-rtx-spark-laptops): The new ProArt P16 and P14 bring NVIDIA's RTX Spark platform into slim creator laptops, promising local agents, large-model workflows, 12K video editing, and MacBook Pro-class ambition.
- [Anthropic ships Opus 4.8 with Dynamic Workflows](https://nowrap.ai/news/claude-opus-4-8): Released May 28, just 41 days after 4.7. The upgrade adds parallel subagent orchestration, effort controls, and benchmark gains in coding and legal work.
- [Google makes Gemini 3.5 Flash the fast path across more AI surfaces](https://nowrap.ai/news/gemini-3-5-flash-default-rollout): Gemini 3.5 Flash is being positioned as Google’s higher-speed default-capable model for apps, APIs, coding loops, and AI-assisted workflows, with broader rollout across Google and partner surfaces.
- [Google introduces Gemini Spark, a 24/7 personal AI agent](https://nowrap.ai/news/gemini-spark-personal-ai-agent): Google is positioning Gemini Spark as a background agent for Gemini Enterprise and Workspace users, designed to carry out recurring tasks, connect to business apps, and work across enterprise workflows with approvals and security controls.
- [Microsoft puts an AI legal agent inside Word for contract review](https://nowrap.ai/news/microsoft-ai-legal-agent-word): Microsoft is bringing a legal-focused AI agent into Word, aiming to help lawyers review contracts, apply playbooks, suggest redlines, and work inside familiar Microsoft 365 compliance controls.
- [Notion is turning its workspace into a hub for AI agents](https://nowrap.ai/news/notion-workspace-hub-ai-agents): Notion is expanding from a document-and-database tool into a broader AI workspace where agents, enterprise search, meeting notes, and team knowledge all live in one operating layer.
- [Google unveils Googlebook, a laptop built around Gemini](https://nowrap.ai/news/googlebook-gemini-intelligence): Google is introducing Googlebook as a new AI-first laptop category, with Gemini-powered contextual actions, generated widgets, deeper Android integration, and premium partner hardware.
- [OpenAI launches Daybreak to push AI deeper into cybersecurity workflows](https://nowrap.ai/news/openai-daybreak-cybersecurity): OpenAI is turning its latest model stack and Codex tooling toward secure code review, threat modeling, patch validation, and defensive cyber operations, with Daybreak positioned as an enterprise security initiative.
- [Grey-market resellers in China are offering cut-rate ChatGPT and Claude access](https://nowrap.ai/news/china-grey-market-chatgpt-claude-api-access): Reporting suggests proxy networks in China are selling discounted access to models like Claude and ChatGPT, raising concerns about prompt harvesting, model substitution, and enterprise data risk.
- [OpenAI launches GPT-Realtime-2 and new voice models for live AI apps](https://nowrap.ai/news/gpt-realtime-2-voice-models-api): OpenAI’s latest realtime stack adds stronger voice reasoning, live translation, and low-latency streaming transcription, pushing voice AI closer to full product infrastructure.
- [Anthropic launches finance agent templates and Microsoft 365 integrations](https://nowrap.ai/news/anthropic-finance-agent-templates-microsoft-365): Anthropic is packaging Claude more aggressively for banking, insurance, and asset-management workflows, with ten ready-made finance agents and tighter integration into Excel, Word, and PowerPoint.
- [Google rolls out Gemini 3.1 Flash Live Preview for voice-first AI](https://nowrap.ai/news/gemini-3-1-flash-live-preview-audio): Google’s latest Gemini live model focuses on audio-to-audio interaction, with more natural speech, stronger instruction-following in longer sessions, and better support for real-time voice agents.
- [xAI rolls out Grok 4.3 with longer context and stronger agent workflows](https://nowrap.ai/news/xai-grok-4-3-launch): xAI says Grok 4.3 brings always-on reasoning, a 1M-token context window, lower pricing, and stronger tool-driven agent behavior, but the launch still lacks the cleaner documentation typical of rival labs.
- [Anthropic pushes Claude deeper into creative work with new tool connectors](https://nowrap.ai/news/claude-creative-work-connectors): Claude is moving beyond general chat and coding into professional creative workflows, with integrations aimed at tools like Adobe Creative Cloud, Blender, Ableton, and Autodesk Fusion.
- [Amazon launches Quick desktop app, an AI assistant that works across files, apps, and team workflows](https://nowrap.ai/news/amazon-quick-desktop-ai-assistant): Amazon is turning Quick into a more ambitious workplace AI product, with a desktop app, persistent work context, shared spaces, workflow automation, and new integrations across common business tools.
- [OpenAI lands on AWS with Bedrock model access, Codex, and managed agents](https://nowrap.ai/news/openai-models-codex-managed-agents-aws): AWS and OpenAI are expanding their partnership to bring OpenAI models, Codex, and Bedrock Managed Agents into Amazon’s enterprise cloud stack, with security and governance as the main selling point.
- [Google pushes Gemma 4 toward local agent workflows with stronger on-device skills](https://nowrap.ai/news/gemma-4-agent-skills): Google is positioning Gemma 4 as a more capable local model for agentic tasks, with an emphasis on edge deployment, developer workflows, and Android-adjacent use cases.
- [Cursor agent powered by Claude deletes PocketOS production database in seconds](https://nowrap.ai/news/cursor-claude-pocketos-db-incident): PocketOS founder Jer Crane says a Cursor workflow running Claude Opus 4.6 wiped production data and volume-level backups in a single Railway API call, turning a staging task into a production incident.
- [OpenAI releases GPT-5.5 and GPT-5.5 Pro, weeks after 5.4](https://nowrap.ai/news/openai-gpt-5-5): A 1M-token context window, sharper agentic coding, and what OpenAI calls its strongest safeguards yet — but Pro tier is six times the price for input.
- [Vercel discloses security incident: employee account compromised via third-party AI tool](https://nowrap.ai/news/vercel-april-2026-incident): A breach of a Context.ai account led to a Vercel employee's Google Workspace, then to plaintext customer environment variables. Rotate now.
- [Anthropic ships Claude Opus 4.7 with stronger coding, agents, and vision](https://nowrap.ai/news/claude-opus-4-7): Released April 16. Early testers report 10–13% improvements on code resolution, with state-of-the-art performance on finance-agent evaluations.

## Glossary
- [AI agent](https://nowrap.ai/glossary/ai-agent): An AI that doesn't just answer questions but takes a goal, makes a plan, and uses tools to carry it out across multiple steps.
- [API](https://nowrap.ai/glossary/api): A doorway that lets one piece of software talk to another — how apps and agents plug into an AI model.
- [Benchmark](https://nowrap.ai/glossary/benchmark): A standardised test used to measure and compare AI models — and a number to read with healthy skepticism.
- [Context window](https://nowrap.ai/glossary/context-window): How much an AI can "hold in mind" at once — the working memory that limits how much it can read or remember in one go.
- [Embedding](https://nowrap.ai/glossary/embedding): A way of turning text into numbers that capture meaning, so an AI can find things by what they mean rather than the words they use.
- [Few-shot prompting](https://nowrap.ai/glossary/few-shot-prompting): Showing an AI a couple of worked examples in your prompt so it copies the pattern instead of guessing.
- [Fine-tuning](https://nowrap.ai/glossary/fine-tuning): Further training a general AI model on your own examples so it adopts a specific style, format, or domain.
- [Foundation model](https://nowrap.ai/glossary/foundation-model): A large, general-purpose AI model trained at huge scale that other tools and products are built on top of.
- [Grounding](https://nowrap.ai/glossary/grounding): Tying an AI's answer to verifiable sources, so it reports what the evidence says instead of what it vaguely recalls.
- [Guardrails](https://nowrap.ai/glossary/guardrails): The rules and checks bolted around an AI to keep its output inside safe, compliant, on-brand bounds.
- [Hallucination](https://nowrap.ai/glossary/hallucination): When an AI states something false with complete confidence — the failure mode that matters most in professional work.
- [Inference](https://nowrap.ai/glossary/inference): The act of running a trained AI model to get an answer — the step you actually pay for and wait on.
- [Knowledge cutoff](https://nowrap.ai/glossary/knowledge-cutoff): The date an AI's training data ends — after which it knows nothing unless you tell it or it can search.
- [Large language model](https://nowrap.ai/glossary/large-language-model): The kind of AI — trained on vast amounts of text — that powers chatbots and writing assistants by predicting the next word.
- [Model Context Protocol](https://nowrap.ai/glossary/model-context-protocol): A common standard that lets AI assistants plug into your tools and data without a custom integration for each one.
- [Multimodal](https://nowrap.ai/glossary/multimodal): An AI that works across more than just text — reading images, audio, and video, not only words.
- [Open-weight model](https://nowrap.ai/glossary/open-weight-model): An AI model whose trained parameters are published, so anyone can download, run, and adapt it themselves.
- [Prompt](https://nowrap.ai/glossary/prompt): The instruction you give an AI — and the single biggest lever you have over the quality of what it gives back.
- [Prompt injection](https://nowrap.ai/glossary/prompt-injection): An attack where hidden instructions buried in a document, email, or webpage hijack an AI into ignoring its real task.
- [Reasoning model](https://nowrap.ai/glossary/reasoning-model): A model that works through a problem step by step before answering, trading speed for accuracy on hard tasks.
- [Retrieval-augmented generation](https://nowrap.ai/glossary/retrieval-augmented-generation): A technique that feeds an AI your own documents at question time, so its answers are grounded in real sources instead of memory.
- [System prompt](https://nowrap.ai/glossary/system-prompt): The standing instructions that set an AI's role and rules before you ever type your first message.
- [Temperature](https://nowrap.ai/glossary/temperature): A setting that controls how predictable or creative an AI's output is — low for consistent, high for varied.
- [Token](https://nowrap.ai/glossary/token): The chunk of text an AI reads and writes in — and the unit you're actually billed by.
- [Tool calling](https://nowrap.ai/glossary/tool-calling): The ability for an AI to use external tools — run a search, send an email, query a system — instead of only producing text.
- [Vector database](https://nowrap.ai/glossary/vector-database): A database built to store embeddings and find items by meaning — the retrieval engine behind most RAG systems.

## For Professionals
- [Teachers](https://nowrap.ai/for/teachers): AI is past the "everyone gets a chatbot" phase and into the boring, useful work — lesson plan iteration, differentiated instruction, parent-newsletter drafts you actually want to send. We list the ones that respect FERPA and student data, and we say so when they don't.
- [Lawyers](https://nowrap.ai/for/lawyers): Contract review at speed without giving up the line you can't cross. Tools that touch privileged work need clear data handling, transparent training policies, and audit trails. We name the ones that have them.
- [Marketers](https://nowrap.ai/for/marketers): Campaign drafts, brand-voice consistency, and competitive research are where AI earns its keep — not the autoplay slop on LinkedIn. We test for actual workflow integration, not demo-day theater.
- [Doctors](https://nowrap.ai/for/doctors): The bar for clinical AI is HIPAA-grade data handling and a clear story on training data. We track scribes, charting copilots, and synthesis tools that practicing clinicians use, and we flag the ones that overstate their evidence.
- [Real Estate](https://nowrap.ai/for/real-estate): Listing copy, market analysis, lead qualification — the unglamorous parts of the job AI is genuinely good at. We track the tools that integrate with MLS data and the ones that just paraphrase your inputs.
- [Freelancers](https://nowrap.ai/for/freelancers): Proposals, invoicing, client communication, project management — the overhead that eats your billable hours. We list tools that respect freelancer pricing and the ones that quietly require a team plan.
- [Engineers](https://nowrap.ai/for/engineers): Coding assistants are a saturated category. We focus on the dimensions that actually matter for production work: privacy of your code, depth of context window, integration with your existing toolchain, and what they cost at team scale.
- [Designers](https://nowrap.ai/for/designers): Image generation got the headlines; the useful tools are about iteration speed and design-system consistency. We track Figma plugins, prototyping copilots, and image tools — and we note which ones you can use in client work without licensing landmines.
- [Writers](https://nowrap.ai/for/writers): Drafting partners, fact-checkers, and editing assistants — not replacement engines. We list the tools that respect a writer's voice and the ones that flatten everything to the median.
- [Researchers](https://nowrap.ai/for/researchers): Literature review and synthesis at the speed of a curious afternoon. We track tools that show their work — citations, source links, and confidence calibration — and we flag the ones that confidently hallucinate.

## Optional
- [About](https://nowrap.ai/about): how Nowrap is edited and verified
- [Submit a tool](https://nowrap.ai/submit): editorial tip line
- [RSS feed](https://nowrap.ai/feed.xml): the Dispatch as RSS 2.0
- [Sitemap](https://nowrap.ai/sitemap.xml): full URL list
- [Full content (llms-full.txt)](https://nowrap.ai/llms-full.txt): every tool, news article, and glossary entry inlined