Devin Desktop makes coding-agent fleet control a buying check
Windsurf is now Devin Desktop, with its product page framing the tool as one surface for managing fleets of local and cloud coding agents. That repositioning lands as Cline adds native macOS and Windows apps, while “I’m looking for an AI coding assistant that can work across a whole repo.” drew 3 searches this week. The decision is shifting from which coding agent writes code to where teams supervise agent work across machines, repositories, and execution environments. Check whether your current coding-agent setup has a credible fleet-management surface before adding another agent.
See what changed on Devin Desktop'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
Gumloop expands its agent platform with Gumball, new models, MCP support… — Gumloop’s updated site highlights Gumball, a proactive model-agnostic self-improving agent, along with new model availability, webhook triggers, Company Brain integrations, external MCP connectivity, agent-managed evaluations, private-clou…
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
|
17/36 · Adaptive Collaborator
Windsurf is rebranded as Devin Desktop, positioning the product as a home… — The product page now presents Devin Desktop as the successor and new name for Windsurf, focused on managing fleets of local and cloud coding agents from one surface.
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
|
14/36 · Adaptive Collaborator
PolyAI launches an agentic dialog platform with self-improving agents… — PolyAI’s redesigned site introduces an enterprise Agentic Dialog Platform, including Wren for continuous conversation analysis, automated improvement recommendations, and deployment monitoring.
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
|
9/36 · Guided Assistant
Klue launches an MCP server and joins Claude’s Connector Directory — Klue has added an MCP server that makes its competitive intelligence available in agent workflows and is now listed in Anthropic’s Claude Connector Directory.
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
|
18/36 · Adaptive Collaborator
Cline adds a native desktop app for macOS and Windows — Cline now presents a native desktop app alongside its IDE and terminal products.
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
|
12/36 · Guided Assistant
Replit Agent adds custom MCP support and expands its integration platform — Replit’s updated Agent experience now supports adding services through custom MCP endpoints, alongside an expanded integrations catalog and new capabilities such as Infinite Canvas and intelligent model routing.
Source: the tool's own site — detected 2026-09-11 by agentic.ai's daily recrawl.
|
What's new in the directory
|
19/36 · Domain Specialist
Atomic Agent is an open-source, local-first AI agent that runs on your own machine and can act through your browser, filesystem, shell, and Git. It uses local models by default, keeps sessions on disk, and pauses for approval on risky actions.
Source: published listing, rubric-scored on ingest.
|
|
19/36 · Domain Specialist
Pragor is a cloud operations layer for teams running AI agents built with frameworks such as CrewAI, LangGraph, and AutoGen. It gives agents and human operators a shared project board with durable context, approval workflows, and an attributable record of work.
Source: published listing, rubric-scored on ingest.
|
|
8/36 · Guided Assistant
HOL Guard sits between AI coding agents and your machine, pausing risky actions for approval before they run. It can intercept shell commands, reads of secret-bearing files, package installs, and MCP server registrations while allowing routine work to continue.
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.
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.
GitHub Copilot helps you write, review, and adapt code directly in GitHub, your IDE, and the terminal. It supports everything from inline suggestions to agentic coding workflows with broader model choices and enterprise controls.
- Cursor · 18/36 · Adaptive Collaborator · L2
Cursor is a developer-focused AI environment that adds agents, context, and automation around your repositories. It combines an editor-like interface, a CLI, and a cloud agent API to automate code review, bug fixing, CI hygiene, and more. Designed for individual developers and engineering teams who want AI to take real actions in their code and infrastructure, not just chat.
- OpenHands · 18/36 · Adaptive Collaborator · L2
OpenHands (formerly OpenDevin) is a fully autonomous open-source AI coding agent. MIT-licensed core, plans + executes Python actions via CodeAct, uses browser + terminal + files. ~72% on SWE-bench Verified with Sonnet 4.5 via the OpenHands Software Agent SDK (arXiv:2511.03690).
What operators are searching for this week
A week's demand signal — what operators are typing into the directory's search. Suggestion-chip defaults + off-topic / greetings filtered.
- "I'm looking for an AI coding assistant that can work across a whole repo."
- "AI agent to triage and reply to my email inbox"
- "best coding agent for a large existing codebase"
- "tools to automate customer support ticket triage"
- "AI agent that can browse the web and fill out forms"
Learn: Action Capability: can it actually DO things (run code, send, edit files), or only suggest?
Action Capability is the part of an agent that turns intent into changes outside the chat box. A tool with high Action Capability can run code, edit files, send messages, create tickets, call APIs, or move data without a human copying each step. A chat interface that drafts an email or writes a script but never executes anything scores much lower, even if it looks polished. This is one clean way to separate a real agent from a chatbot with extra UI. Check the tool’s action surface: does it show concrete actions, logs, permissions, and success or failure states for each step? If you can only get text suggestions, it’s not doing the work yet.
How our 9-dimension rubric works →
One tool worth testing
Atomic Agent — Atomic Agent is a local-first open-source agent for browser, filesystem, shell, and Git work on an operator’s own machine. It is worth testing now because the new 19/36 listing pairs local models and on-disk sessions with explicit approval pauses for risky actions, a useful counterpoint to cloud-managed coding-agent fleets. Evaluate it against your current coding agent in a 30-minute pilot: complete one repository task, trigger a shell or file-risk action, and inspect whether its approval boundary fits your team’s controls.
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 →
|