Releases·Tool Releasefeatured
Anthropic launches a portal to submit and track Claude plugins as MCP usage grows 110x this year
Anthropic announced a new developer portal for building and submitting plugins for Claude. Plugins package MCP servers and skills, and Anthropic says they are becoming the primary way to build on the platform. Submissions are made directly in Claude, with usage analytics provided once plugins go live.
3374 engagement4 sourcesFri, Sep 25, 2026, 08:06 PM
AI Overview
- Anthropic (X, 2026-09-25) launched a portal for developers to submit plugins, track review status, and see usage.
- Plugins package MCP servers and skills, and Anthropic positions them as the way to build for Claude.
- Anthropic says MCP usage across Claude products is up 110x this year.
- Anthropic (X, 2026-09-25) says submissions go through two paths: a single remote MCP connector or a plugin bundle repo combining MCP servers and skills, made directly in Claude.
- Anthropic (X, 2026-09-25) says usage analytics show installs by product and version, listing views, and which searches lead to the listing.
Related
Social·Community
Two Google accounts in Claude's Gmail/Drive connectors?
A user with both a personal Gmail and a Workspace account reports that Anthropic's Claude Gmail/Drive connectors can only be linked to one Google account per service, making account switching impractical. Workarounds via Google's official MCP server as a custom connector fail, and the user asks whether multi-account support is planned.
A user with both a personal Gmail and a Workspace account reports that Anthropic's Claude Gmail/Drive connectors can only be linked to one Google account per service, making account switching impractical. Workarounds via Google's official MCP server as a custom connector fail, and the user asks whether multi-account support is planned.
21 engagement1 sourceredditSat, Sep 26, 2026, 05:40 AM
Products·launched
Read-only Google Ads MCP connector for Claude that surfaces wasted spend
A Google Ads MCP connector built with Claude Code that lets Claude read an ads account read-only to find search terms with spend but no conversions. Because Google's OAuth scope grants full edit access, the tool never calls mutate endpoints; instead it outputs a negative keywords file for review in Google Ads Editor. It targets advertisers who want AI analysis without giving an LLM control over live campaigns.
A Google Ads MCP connector built with Claude Code that lets Claude read an ads account read-only to find search terms with spend but no conversions. Because Google's OAuth scope grants full edit access, the tool never calls mutate endpoints; instead it outputs a negative keywords file for review in Google Ads Editor. It targets advertisers who want AI analysis without giving an LLM control over live campaigns.
18 engagement2 sourcesredditWed, Sep 23, 2026, 09:35 AM
Products·building
CapForge: Local video captioning app with Claude MCP integration
CapForge is a local desktop app that transcribes short-form video and burns animated captions. It integrates with Claude via MCP, allowing the AI to read transcripts, edit captions, and render frames for visual feedback. It solves privacy and cost issues for creators who don't want to upload unreleased footage to cloud services.
CapForge is a local desktop app that transcribes short-form video and burns animated captions. It integrates with Claude via MCP, allowing the AI to read transcripts, edit captions, and render frames for visual feedback. It solves privacy and cost issues for creators who don't want to upload unreleased footage to cloud services.
493 engagement1 sourceredditWed, Aug 5, 2026, 09:03 AM
Social·Community
Anthropic's MHS preview standardizes lab-equipment control, but Claude retries caused physical errors in experiments
A Reddit discussion examines Anthropic's MHS preview, which gives programmable lab and manufacturing equipment a shared driver format reachable via MCP, CLI, or code. The poster argues that standardizing device access does not make agents safe to drop into physical control loops, citing a case where Claude retried a failed physical operation and made the problem worse.
A Reddit discussion examines Anthropic's MHS preview, which gives programmable lab and manufacturing equipment a shared driver format reachable via MCP, CLI, or code. The poster argues that standardizing device access does not make agents safe to drop into physical control loops, citing a case where Claude retried a failed physical operation and made the problem worse.
6 engagement1 sourceredditThu, Sep 3, 2026, 05:14 PM
Products·launched
Automated MCP server security scoring index
A trust index that automatically scores MCP servers from the official registry based on security criteria like runtime guardrails and SAST scans. It helps security professionals quickly assess whether an MCP server is safe for enterprise use, reducing manual review bottlenecks.
A trust index that automatically scores MCP servers from the official registry based on security criteria like runtime guardrails and SAST scans. It helps security professionals quickly assess whether an MCP server is safe for enterprise use, reducing manual review bottlenecks.
1 engagement1 sourcehackernewsSat, Jul 11, 2026, 06:57 PM