Cursor’s AI teammates change what you must sandbox
Cursor’s new “AI teammates” can sign in to users’ tools, operate them, and return completed work—the missing detail behind its new homepage claim. It also leads outbound clicks this week, suggesting operators are evaluating coding agents as actors inside existing systems rather than chat interfaces. ThunderPhone 2.0 makes that same operational shift self-serve for voice agents, with guided SIP setup and production call monitoring, while Tradehand now documents both a public OpenAPI surface and MCP endpoint. Check whether each tool’s identity, approval, and audit controls match the systems you would let it touch.
See what changed on Cursor's listing →
What changed on the ground
Changes we detected on each tool's own pages — pricing, features, platform reach — via our daily crawls. Often ahead of any press coverage; each links to the scored listing.
|
13/36 · Adaptive Collaborator
Phrase announces major platform release with conversational operations, MCP… — Phrase’s September 2026 release expands its Language Intelligence Platform with conversational platform operations, an MCP Server, new APIs for Style Guides, Quality Profiles and Language Assets, Phrase Studio for multimedia localization,…
Source: the tool's own site — detected 2026-09-26 by agentic.ai's daily recrawl.
|
|
18/36 · Adaptive Collaborator
ThunderPhone 2.0 launches self-serve voice AI platform with new pricing… — ThunderPhone has launched ThunderPhone 2.0, introducing self-serve access alongside expanded voice-agent capabilities, including MCP and knowledge-base integrations, guided SIP setup, automated agent improvement, and production call monito…
Source: the tool's own site — detected 2026-09-27 by agentic.ai's daily recrawl.
|
|
13/36 · Adaptive Collaborator
Tradehand expands its trades marketplace with public API and MCP support — Tradehand’s updated site documents a broader agentic platform for UK trades businesses, including a public OpenAPI surface, API catalog, MCP server card, and MCP Streamable HTTP endpoint.
Source: the tool's own site — detected 2026-09-27 by agentic.ai's daily recrawl.
|
|
18/36 · Adaptive Collaborator
Cursor adds AI teammates that can operate tools and return completed work — Cursor’s homepage now promotes “AI teammates” that can sign in to users’ tools, operate them, and return finished work.
Source: the tool's own site — detected 2026-09-26 by agentic.ai's daily recrawl.
|
|
10/36 · Guided Assistant
Dust updates pricing and expands its enterprise AI agent platform — Dust has introduced revised self-serve and enterprise pricing, including credit-based plans, workspace-pooled credits, volume pricing, and expanded enterprise controls.
Source: the tool's own site — detected 2026-09-27 by agentic.ai's daily recrawl.
|
|
13/36 · Adaptive Collaborator
Gumloop adds password-protected artifacts, agent voice conversations… — Gumloop's updated product site highlights several concrete launches, including password-protected artifact sharing, faster and cheaper agent tasks, voice conversations with agents, browser notifications for agent runs, and Gmail and Linear…
Source: the tool's own site — detected 2026-09-26 by agentic.ai's daily recrawl.
|
What's new in the directory
|
2/36 · Reactive Tool
WSUP AI is a browser-based AI character chat service for story-driven conversations, roleplay, and virtual companionship. You can start chatting with its catalog of themed characters without creating an account.
Source: published listing, rubric-scored on ingest.
|
|
4/36 · Reactive Tool
ForgeSprite is a browser-based AI generator for static 2D game assets such as characters, props, inventory items, skill icons, scenes, and textures. You describe an asset, optionally add a reference image, review a generated sample and credit quote, then export PNGs or arrange completed frames into a sprite sheet.
Source: published listing, rubric-scored on ingest.
|
Agenticness Leaderboard
This week's most-clicked tools, re-ranked by rubric score. The score is the editorial signal; the popularity is what drew the directory's traffic here this week.
- Hyper · 22/36 · Domain Specialist · L3
Hyper is a cloud-based autonomous software engineering platform for turning product objectives into tested code in your GitHub repositories. It plans work, runs parallel agent sessions in isolated lanes, verifies results server-side, and can open pull requests or deploy supported projects.
- Sinatra · 21/36 · Domain Specialist · L3
Sinatra is an autonomous coding agent that takes work from Linear or GitHub, reads the relevant repository, writes a change, runs tests, and opens a draft pull request. Teams review the result in GitHub and can request follow-up changes in the PR thread.
Claude Code is Anthropic's agentic coding tool that lives in your terminal. It understands your entire codebase, makes multi-file edits, runs commands, manages git workflows, and uses MCP for tool integration. Built with a Unix philosophy — it reads, plans, edits, and verifies in a loop. The fastest-growing product in the coding agent category.
- Cline · 18/36 · Adaptive Collaborator · L2
Cline is an AI coding assistant for VS Code that can inspect your project, edit files, run terminal commands, and use a browser while asking for permission at each step. It is aimed at developers working on real codebases who want more than code completion.
Learn: Adaptation: when a step fails, does it re-plan - or just stop?
Adaptation is how a tool responds when an action fails, a file is missing, an API call errors, or the first plan stops making progress. A weak system just surfaces the error or asks the user to try again. A stronger agent notices the failure, changes the next step, and keeps working toward the goal. That gap separates a real agent from a chatbot with a nice wrapper. One concrete check: look for an example where the tool hits a dead end and then chooses a different path on its own. If every failure ends in a human prompt, the system may be helpful, but it isn’t adapting.
How our 9-dimension rubric works →
One tool worth testing
ThunderPhone — Test ThunderPhone for a self-serve voice agent handling a bounded inbound-call workflow. ThunderPhone 2.0 is the test now because it pairs guided SIP setup and knowledge-base/MCP integrations with production call monitoring, moving the service closer to an operated phone workflow rather than a voice demo. In a 30-minute pilot, route one low-risk call type through ThunderPhone and compare its handoff, escalation, and monitoring controls against your current phone or contact-center path.
Our pick: ChatGPT Deep Research
The use case: research & deep analysis.
Tops our independent /36 — strongest on autonomy and planning, ahead of Gemini Deep Research. Pricing: Subscription (quotas undisclosed).
Ranked by our 9-dimension Agenticness rubric — not by who pays us (nobody does). The full scorecard, true-cost pricing, and the reliability evidence behind every score are in the report.
Read the full report →
|