News & Updates
June 9, 2026
Phoenix function disassembly and applicable scenario analysis
Phoenix helps teams establish a closed loop for Agent observation and evaluation, supports local, Docker, Kubernetes and cloud deployment, and is suitable for AI product teams that continue to iterate.
Source: Phoenix official documentation
Perplexity Product Observation: Capability Inventory in AI Search Direction
Perplexity is an AI-native search/answer engine founded by Aravind Srinivas and others, providing Web, iOS, Android, desktop, Comet browser and Sonar API.
Source: Perplexity official documentation
Phind function dismantling and applicable scenario analysis
Phind is an AI search engine designed for developers. It combines real-time web search with LLM reasoning to generate technical answers with code examples. It supports Web, API and VS Code plug-ins. The free version is available, and the Pro version is $17/month.
Source: Phind official documentation
How to use Otter.ai: the leading AI meeting transcription and note-taking assistant
Otter.ai is an AI meeting transcription and note-taking assistant that transcribes Zoom, Teams, and Google Meet meetings in real time, automatically generates summaries and action items, and supports conversational Q&A. Free version 300 minutes/month, Pro version $16.99/month.
Source: Otter.ai official documentation
Understand Originality.AI in one article: Designed for content marketing and SEO teams
Originality.AI is an AI detection and plagiarism detection platform specially designed for content marketing and SEO teams. Founded by a Canadian team, it supports GPT-4, Claude, and Gemini output recognition, and provides readability scoring and team collaboration functions.
Source: Originality.AI official documentation
Opik function dismantling and applicable scenario analysis
Opik is suitable for Agent and LLM application teams, providing tracking, evaluation, optimization and production monitoring, and supports Open Source and Cloud solutions.
Source: Opik official documentation
Re-understand Opus Clip: the core capabilities of AI video tools
Opus Clip uses AI to automatically extract the most viral clips from long videos, automatically adds subtitles, reconstructs the scene to 9:16, and quickly generates short videos suitable for TikTok, Reels, and YouTube Shorts. It is used by 1M+ users.
Source: Opus Clip official documentation
OpenRouter function dismantling and applicable scenario analysis
This article dismantles OpenRouter from parameters, functions, costs, versions and implementation boundaries to help the team form executable judgments during the pilot and expansion stages.
Source: OpenRouter official documentation
Obviously AI function dismantling and applicable scenario analysis
Obviously AI is a code-free predictive analysis platform that automatically completes feature engineering, model training and prediction deployment by uploading a CSV file. It supports customer churn prediction, sales prediction and other business scenarios. It is targeted at non-technical business teams, and pricing starts from $75/month.
Source: Obviously AI official documentation
Ollama Product Observation: Capability Inventory in AI Model Training Direction
Ollama is an open source local large model running tool with 173K+ stars on GitHub. It can run Llama, DeepSeek, Qwen and other models on macOS/Windows/Linux with one line of commands.
Source: Ollama official documentation
Notion AI Product Observation: Capability Inventory in the Direction of AI Office and Efficiency
Notion AI deeply embeds LLM capabilities into team documents and databases, providing agents, cross-application search, meeting records, and process automation. This article summarizes its capability stack, pricing and applicable scenarios.
Source: Notion AI official documentation
Re-understand n8n: the core capabilities of AI office and efficiency tools
n8n is a self-hosted open source workflow automation platform that supports AI Agent nodes. This article summarizes its functions, pricing and application scenarios.
Source: n8n official documentation
Understand NotebookLM in one article: Google’s research and writing based on Gemini
NotebookLM is Google's Gemini-based AI research and note-taking assistant that integrates user documents, web pages, and videos into a conversational knowledge base. This article summarizes its functions, pricing and scenarios.
Source: NotebookLM official documentation
Understand Murf AI in one article: 120+ AI sounds, 20+ languages
Murf AI provides 120+ realistic AI voices and 20+ languages, supports voice cloning, video dubbing synchronization and API integration, and is widely used in eLearning, marketing videos and audiobook production. Creator starts at $29/month.
Source: Murf AI official documentation
What is Mistral Le Chat: European-native AI by Mistral AI
Mistral Le Chat is the flagship conversation assistant of French Mistral AI, providing capabilities such as Research, Code, Canvas and image generation. This article summarizes its capabilities, pricing and European compliance advantages.
Source: Mistral Le Chat official documentation
How to use Modal: Use Python decorator to run functions on the cloud with one click
Modal is a serverless GPU cloud platform driven by Python decorators. Developers only need to add decorators such as @app.function(gpu="A100") to Python functions to run local code on cloud GPUs with one click.
Source: Modal official documentation
July 9, 2026
Meta releases Muse Spark 1.1: personal super-intelligent model base upgraded
Meta released Muse Spark 1.1 through the Meta model API on July 9, which is an upgraded version of its flagship model in the direction of "personal super intelligence"; the first version was released on April 8, and together with Muse Image and Muse Video released on July 7, it forms the Meta multi-modal generation matrix, continuing the "open source + cutting-edge" dual-track strategy.
Source: Meta AI official blog
June 30, 2026
MiniMax Speech 2.8 and Music 3.0: Dual-line upgrade of speech and music generation
MiniMax released the Speech 2.8 voice model and Music 3.0 music model in June, both providing APIs through platform.minimax.io; together with M3 and H3, they form a multi-modal matrix, showing its dual-line commercial layout in C-side entertainment and B-side APIs.
Source: MiniMax official
June 1, 2026
MiniMax M3: "All in One" flagship with cutting-edge coding, 1M context, and native multi-modality
MiniMax released its flagship model M3 in May, which is based on the new MSA sparse attention architecture and supports 1M token context and native multi-modality. It is equipped with MiniMax Code, Hub and Audio, positioning it as a universal smart base that can undertake long-range and complex tasks.
Source: MiniMax official blog
August 2, 2026
MiniMax H3 officially released: unified generation of all modes, 15 seconds of 2K video and 1/3 the price
MiniMax H3 is officially released. It is a universal full-modal generation model that supports unified understanding and output of text/image/video/sound. It can generate up to 15 seconds of 2K video. The price is about 1/3 of the mainstream model.
Source: user feed