Signals, labelled honestly
AI News and Model Release Coverage
AI launch news has a verification problem: a screenshot or a post on X can travel further in six hours than an official model card does in six days. AIXList's research hub covers launches with an explicit line between confirmed announcements and reported claims, and this page indexes that coverage alongside the tools we catalogued as a result. Use it as a reading order for a fast-moving release — not as a feed of everything trending, and not as a substitute for the vendor's own documentation.
Curated tools
18 picks from the AIXList catalogue. Every card links to a full listing with features, use cases and FAQ.
AXL
AXL (axl.tech) is an AI-native go-to-market platform for AI product founders that folds websites and funnels, CRM pipelines, marketing automation, email, AI chat and vo…
ProductivityFreemiumtoolNoMac
NoMac (nomac.app) rents real cloud Macs to AI agents by the minute: full macOS on Apple silicon with Xcode preinstalled, admin/root access, and iOS Simulator, driven ov…
Code & DevPay as you gotoolTobira
Tobira (tobira.ai) turns your website into an AI site agent that sells to both humans and AI agents: paste your URL and it scans your services, pricing, and blog in min…
ProductivityFreemiumtoolTurnstone
Turnstone (myturnstone.ai) is a YC-backed Mac app for personal AI agent teams that is model-agnostic by design: agents run on the ChatGPT, Claude, Grok, or Cursor subsc…
ProductivityFreemiumtoolOvermind
Overmind (overmindlab.ai) is an open platform for agent continuous learning: it captures your AI agents' production traces and trains team-owned specialist models from…
Code & DevOpen SourcetoolSpacesheep
Spacesheep (spacesheep.dev) is a Y Combinator F26 startup that turns AI agent outputs into collaborative interactive spaces—docs, boards, and prototypes agents and team…
ProductivityFreemiumtoolEntire
Entire (entire.com), founded by former GitHub CEO Thomas Dohmke, is an agent-native Git forge where Trails replace pull requests—pairing self-hosted Git infrastructure…
Code & DevFreemiumtoolParadev
Paradev (paradev.com)—formerly Superset IDE—is a parallel coding-agent IDE that relaunched under a new name in October 2026 alongside an announced $11.5M seed led by US…
Code & DevFreemiumtoolPomerado
Pomerado (YC F26, pomerado.com) turns any website into API and MCP tools that AI agents call directly—claiming 2-4x faster responses at 10-1000x lower cost than browser…
Code & DevFreemiummcpHallmark
Hallmark by Nutlope (github.com/Nutlope/hallmark) is an open-source design skill for Claude Code, Cursor, and Codex that checks generated UI against anti-AI-slop design…
Code & DevFreetoolHyperFrames Studio by HeyGen
HyperFrames Studio by HeyGen (hyperframes.heygen.com/studio-app) is a desktop app for agent-driven video: coding agents author HTML animations, and Studio renders them…
Video GenerationFreetoolBits
Bits (bits.interfacesforhuman.com) is an open-source personal agent that runs locally on your machine, keeping its memory in plain Markdown files and producing WebMCP a…
ProductivityFreetoolCauso
Causo (causo.ai) is AI lead research for sales teams: it hunts down prospects who do not appear in your database and delivers findings through an MCP server, so agents…
Marketing & SEOFreemiumtoolCloudKnife
CloudKnife (cloudknife.com) runs AI agents that learn how your Azure resources are actually used, then applies deterministic rules to trim the waste—cutting spend witho…
Code & DevFreemiumtoolComplete AI Agents
Complete AI Agents (completeaiagents.com) is a library of 9,866 MCP-built agent templates covering 495 jobs across 42 industries, per its October 17, 2026 announcement—…
ProductivityFreemiumtoolCrawlForge MCP
CrawlForge MCP (crawlforge.dev) packs 31 specialized MCP tools into one server—scraping, crawling, web search, deep research, and autonomous extraction—so an agent gets…
Code & DevFreemiummcpDadan
Dadan (dadan.io) is an AI video workspace that puts screen recording, meeting capture, and tutorials in one place, with AI-assisted editing, notes, and summaries layere…
Video GenerationFreemiumtoolDocuMemory
DocuMemory (documemory.dev) distills your Claude Code, Cursor, and ChatGPT sessions into searchable decision notes, then serves them back through MCP—so the reasoning b…
Code & DevPaidmcp
Confirmed, reported, or speculation
Every research article states the status of its claims. Confirmed means an official announcement, documentation or model card exists. Reported means credible posts or press coverage describe a capability the vendor has not published. Speculation is labelled and kept out of capability tables. This convention exists because planning a workflow around an unreleased feature is expensive, and the cost lands on the reader rather than on whoever posted the rumour.
From news to a usable listing
When a launch results in a real product with a working URL and a describable use case, it is added to the directory as a tool, agent or MCP server. The recently added page shows those listings in chronological order. Model releases that do not have a standalone product page are covered in the research hub instead, so the directory stays a list of things you can actually use.
How to keep up without noise
Pick one or two categories that matter to your work and follow those category pages plus the research articles tagged to them. Broad AI news feeds are mostly irrelevant to any single workflow, and the highest-signal change is usually a pricing or API update on a tool you already use rather than a brand-new model announcement.
Go deeper
Frequently asked questions
- Is this a live news feed?
- No. It is an index of AIXList research articles and the tool listings derived from them. Articles are dated, and their status labels tell you how current the underlying claims are.
- What does 'reported' mean on AIXList?
- A capability or release described in posts or press coverage but not confirmed by the vendor in official documentation. Articles label these claims explicitly and avoid presenting them as specifications.
- Do you cover every AI model release?
- No. Coverage is selective and focused on launches relevant to the directory's categories — tools, agents, MCP servers, and image, video, voice and coding models.
- How do I suggest a correction?
- Use the contact link in the footer with the article URL and a source. Corrections with primary-source documentation are prioritised.