Skip to main content
Back

Google lets AI agents make payments

New agent payment protocol, Alibaba web agent, GitHub MCP directory, Tencent 3D model, Cursor commands

September 17, 2025 4 min read

Hey

Welcome to AlphaSignal, the most read source of news by AI engineers and researchers.


Every day, we identify and summarize the top 1% of news, papers, models, and repos, so you’re always up to date.


Here’s today’s roundup:

Together with:

alpha_signal_image_2

Summary

Read time: 4 min 25 sec

Top News

▸ Google’s new protocol lets AI agents securely make payments for you

Assembly AI

▸ Build scalable Conversational AI with AssemblyAI’s accurate speech-to-text

Top Signals

▸ Alibaba unveils the first fully open-source web agent rivaling OpenAI’s

▸ GitHub provides curated MCP directory with one-click installation

▸ Tencent’s new 3D generation model improves realism, textures, and 3D accuracy

▸ Cursor enables custom slash commands and context summarization

▸ Sam Altman outlines stricter teen safeguards, age checks, parental controls

micro1

▸ Micro1's AI recruiter agent hits $500M valuation and 40% growth

Trending Repos

▸ nocodb: provides open-source Airtable alternative for database management

▸ Real-Time-Voice-Cloning: Clone voices instantly from seconds of audio input

▸ n8n-workflows: Library of 2,053 ready-to-use automation workflows

Top Lecture

▸ Stanford CS231n this year: adds video, 3D vision, and robot learning

Top News

Google launches AP2, enabling AI agents to securely authorize and complete payments across platforms

7,942 Likes

alpha_signal_image_3

Google launched the Agent Payments Protocol (AP2), an open framework that enables AI agents to securely authenticate, authorize, and complete payments across platforms.


Backed by over 60 financial and technology organizations, AP2 extends the Agent2Agent (A2A) and Model Context Protocol (MCP).


Key Features

  • Mandates: Signed digital contracts prove what the user authorized.

  • Dual approval: Users approve both intent (search) and cart (final items).

  • Delegated tasks: Pre-signed Intent Mandates allow agents to execute purchases automatically under user-defined conditions.

  • Universal payments: Works with cards, bank transfers, and crypto (stablecoins, web3).

  • Audit trail: Every step is logged for security and accountability.

How It Works

  • Intent Mandate: User tells an agent what to buy (e.g., shoes, tickets). This request is recorded as a verifiable contract.

  • Cart Mandate: Agent presents options. User signs off on the final selection, price, and items.

  • Delegated Mode: For tasks without the user present, the Intent Mandate already defines rules. Agent executes a Cart Mandate automatically when conditions are met.

  • Payment Link: The chosen method (card, bank, or crypto) connects securely to the Cart Mandate.

  • Audit Trail: The full chain, intent, cart, payment is cryptographically secured and reviewable.

Industry Support

More than 60 organizations including American Express, Coinbase, Mastercard, PayPal, Revolut, Salesforce, and UnionPay are collaborating to shape AP2.


Availability

  • Technical specifications and reference implementations are available on GitHub.

  • The framework supports both traditional payments and new crypto integrations (via the A2A x402 extension with Coinbase, Ethereum Foundation, and MetaMask).

  • Industry partners will deploy AP2-enabled agents through Google’s AI Agent Marketplace and Google Cloud Marketplace.

Try now
alpha_signal_image_4

Startups Choose AssemblyAI for Reliable, Affordable Speech-to-Text

Startups like Granola and Cluely rely on AssemblyAI to power speech-to-text inside their products. With high accuracy, low latency transcription, and pricing starting at $0.15/hr, AssemblyAI is the foundation for building Conversational AI apps that scale.


With AssemblyAI, you can:

  • Transcribe speech with 93%+ accuracy across 99+ languages

  • Build low-latency Conversational AI apps that scale seamlessly

  • Keep costs predictable with pricing from $0.15/hr

  • Integrate easily with developer-friendly APIs and clear documentation

Start building scalable Conversational AI products with AssemblyAI today.

Get started free

partner with us

Signals

Alibaba launches Tongyi DeepResearch, an open-source web agent rivaling OpenAI Deep Research with only 30B parameters

2,847 Likes

GitHub introduces MCP Registry as a central hub for discovering trusted, open-source servers

1,258 Likes

Tencent upgrades Hunyuan3D, its 3D generation model, enhancing realism, textures, and facial contours

1,135 Likes

Cursor adds custom commands, summarization, MCP Resources, and improved terminal

1,074 Likes

Sam Altman explains OpenAI’s approach to balancing privacy, freedom, and teen safety

7,561 Likes

micro1 hits $500M valuation and 40% growth powering frontier AI labs

micro1, valued at $500M after a $35M Series A, connects top experts, from PhDs to doctors and engineers, with frontier AI labs and Fortune 10 companies.

Its AI recruiter agent, Zara, has powered 500K+ AI-led interviews, fueling high-quality human data for AI’s toughest problems.

Learn more ↗️

Repos

nocodb

57,027 Stars

This repo offers an open-source Airtable-alternative that auto-generates REST APIs over spreadsheets or SQL databases, enabling data modelling, role-based access, and integrations for internal tools.

Real-Time-Voice-Cloning

56,243 Stars

It provides pre-trained models for speaker encoding, synthesizer, and vocoder. You can generate natural multi-speaker speech and experiment with state-of-the-art voice cloning pipelines locally.

n8n-workflows

31,291 Stars

This repo provides 2,053 automation workflows for n8n, covering integrations, triggers, and complex tasks. It helps you import, adapt, and deploy workflows to accelerate automation projects.

Top Lecture

Stanford CS231n this year: adds video, 3D vision, and robot learning

2,849 Likes

alpha_signal_image_5

This year’s Stanford Spring Deep Learning for Computer Vision (CS231n) opens with Fei-Fei Li and Ehsan Adeli on the core of modern vision: learning hierarchical image representations and solving real tasks.


The 10-week course covers:

  • Core tasks: image classification, object detection, and segmentation

  • Architectures: CNNs, recurrent nets, transformers, diffusion models, vision-language models

  • Advanced topics: video understanding, 3D vision, self-supervised learning, generative models, robot learning

  • Human-centered AI and societal impact

The course emphasizes both foundational deep learning theory and SOTA vision methods used in search, medicine, mapping, drones, and autonomous driving.

Watch now

At Alpha Signal, our mission is to build a sharp, engaged community focused on AI, machine learning, and cutting-edge language models, helping over 200,000 developers stay informed and ahead. We’re passionate about curating the best in AI, from top research and trending technical blogs to expert insights and tailored job opportunities. We keep you connected to the breakthroughs and discussions that matter, so you can stay in the loop without endless searching. We also work closely with partners who value the future of AI, including employers and advertisers who want to reach an audience as passionate about AI as we are.


Our partnerships are based on shared values of ethics, responsibility, and a commitment to building a better world through technology.Privacy is a priority at Alpha Signal. Our Privacy Policy clearly explains how we collect, store, and use your personal and non-personal information. By using our website, you accept these terms, which you can review on our website. This policy applies across all Alpha Signal pages, outlining your rights and how to contact us if you want to adjust the use of your information. We’re based in the United States. By using our site, you agree to be governed by U.S. laws.

Looking to promote your company, product, service, or event to 250,000+ AI developers?

Work with us