Unreleased AI Features - TestingCatalog (Page 12)

8 Oct 2025 · 2 min read

Google tests MCP support for Opal workflow builder

For Opal’s audience, MCP support would enable integrations with third-party services and custom tools, directly in workflows.

8 Oct 2025 · 2 min read

Anthropic prepares Claude Code release for mobile apps

Anthropic prepares a Code section on web and mobile with GitHub integration, repository browsing, and Claude Code tasks tailored to developers.

7 Oct 2025 · 2 min read

Perplexity to launch GPT-5 Pro for Max subscribers

Perplexity offers GPT-5 Pro access to Max subscribers, pairing OpenAI’s top-tier model with higher compute usage and premium performance.

7 Oct 2025 · 2 min read

XAI readies Grok tools support in Tasks for data fetching

xAI prepares Grok web tools to pull data from Gmail, Slack, Notion and X search while providing ticker stock updates and unified insights.

7 Oct 2025 · 1 min read

Google NotebookLM Magic View to appear on main chat widget

Google NotebookLM's Magic View now appears atop the chat interface, fueling speculation over its generative visuals amid an unclear purpose.

7 Oct 2025 · 1 min read

Google works on annotation feature for AI Studio apps builder

Google upgrades AI Studio in Apps Builder with a dictation tool and tests UI annotations to aid swift app coding and troubleshooting.

6 Oct 2025 · 2 min read

OpenAI may release new Image model and Guardrails soon

OpenAI tests an image model selector with varied outputs along with a new Guardrails dashboard for API safety, targeting developers and businesses.

5 Oct 2025 · 2 min read

OpenAI prepares to release Agent Builder during DevDay on October 6

Agent builder will let users build their agentic workflows, connect MCPs, ChatKit widgets and other tools. This is one of the smoothest Agent builder canvases I've used so far.

5 Oct 2025 · 3 min read

Google A/B tests Gemini 3 on AI Studio ahead of upcoming release

Google is likely to release Gemini 3 on October 9 after extensive testing on AI Studio. Early tests show significant performance on coding tasks.

4 Oct 2025 · 2 min read

Google tests fully autonomous plan in Jules agent

Google’s Jules tests a fully autonomous plan as the agent automates complex coding tasks from branch creation to pull request merges.

3 Oct 2025 · 1 min read

NotebookLM Video Overviews will get new styles, including Anime

What do we know so far? Video Overviews on NotebookLM are likely to get an upgrade soon. New Formats and visual styles will make them more visually compelling.

2 Oct 2025 · 2 min read

OpenAI works on user profiles and direct messaging for ChatGPT

What do we know so far? OpenAI is likely preparing ChatGPT to become more "Social" so users will be able to customise their profiles and also send direct messages to others via the same app.

1 Oct 2025 · 2 min read

Microsoft develops Copilot Connectors, Coco mode and Email Assistant

Connectors are coming to Copilot! Also, Coco mode is a new chat option labelled as "Warm and intuitive" where users will be able to "Chat with Coco about life stuff". Copilot therapy 👀

1 Oct 2025 · 1 min read

Anthropic works on customizable Skills for Claude

Anthropic works on Skills Preview in Claude settings, letting users upload custom skill files for tailored automation.

30 Sep 2025 · 1 min read

Google develops Infographics tool for NotebookLM

Google develops a NotebookLM infographics artifact that lets users choose language and focus to turn dense notes into shareable graphics.

29 Sep 2025 · 2 min read

Hume AI tests Octave 2 Multilingual text-to-speech model

Hume AI is testing Octave 2 Multilingual internally, a text-to-speech model with low-latency voice synthesis across 10+ languages for real-time audio.

28 Sep 2025 · 1 min read

Google readies NotebookLM persistent chat history

Google readies NotebookLM update with persistent chat history, allowing users to retrieve old chats for continuous research and teamwork.

27 Sep 2025 · 6 min read

Anthropic experiments with an agent for real-time UI generation

What do we know so far? "Imagine with Claude" will be released as a temporary demo for certain plans (only Max?). Users will be interacting with a classic desktop UI where windows and apps are managed by the AI itself.