{"data":{"agents":[{"agent":{"id":"aider","name":"Aider","aliases":["aider-chat","aider AI","aider pair programmer"],"provider":"Paul Gauthier (open source)","website":"https://aider.chat","supported_ide":["Terminal / CLI"],"updated_at":"2026-09-30","description":"Aider is a free, open-source AI pair programming tool that runs in your terminal. It works directly with your git repository, supports all major LLM providers (Anthropic, OpenAI, Google, Ollama, and more), and specializes in multi-file edits with automatic git commits.","faq":[{"question":"What is Aider and how does it work?","answer":"Aider is a command-line tool that lets you pair program with an LLM inside your terminal. You add files to the chat context, describe what you want, and Aider applies the changes directly to your code with an automatic git commit. It works with any git repository."},{"question":"Which AI models does Aider support?","answer":"Aider can connect to almost any LLM, including local models. Recent development added support for Claude Sonnet 4.6 and Claude Opus 4.7 (including Opus 4.7 via Bedrock, Vertex, and OpenRouter) and GPT-5.5, and maps the opus and sonnet aliases to the latest Claude models. You specify the model with the --model flag."},{"question":"Is Aider free to use?","answer":"Aider itself is completely free and open-source (Apache 2.0). You pay only for the API usage of the underlying LLM provider you choose (e.g., Anthropic, OpenAI). Using Aider with local models via Ollama incurs no API costs at all."},{"question":"Does Aider require a specific IDE?","answer":"No. Aider runs entirely in the terminal and is IDE-agnostic. You can use it alongside any editor — VS Code, Vim, Emacs, JetBrains IDEs, etc. Some editors have community-built Aider integrations, but none are required."},{"question":"How does Aider handle multi-file edits?","answer":"Aider uses a repo-map system that builds a high-level map of your entire codebase (classes, functions, signatures) to help the LLM understand the project structure. You add specific files to the chat, and Aider applies coordinated changes across all of them in a single operation."},{"question":"Does Aider commit changes automatically?","answer":"Yes. By default, Aider creates a git commit after each successful change, with a descriptive commit message generated by the LLM. You can disable this behavior with --no-auto-commits if you prefer to review changes before committing."},{"question":"What is Aider's Architect mode?","answer":"Architect mode separates planning from execution: a powerful model (e.g., Claude Opus 4.6) plans the changes, then a faster/cheaper model (e.g., Claude Haiku 4.5) executes the edits. This improves quality while reducing cost for large tasks."},{"question":"Can Aider be used offline or with local models?","answer":"Yes. Aider supports local LLMs via Ollama, LM Studio, and other OpenAI-compatible local servers. This enables fully offline use with no external API calls or costs."},{"question":"How does Aider compare to Claude Code?","answer":"Both are CLI-based coding agents. Claude Code is a proprietary tool from Anthropic, tightly integrated with Claude models only. Aider is open-source and model-agnostic, supporting dozens of LLMs. Aider is free (API costs aside), while Claude Code requires a paid Claude subscription or an Anthropic Console account."},{"question":"Does Aider support coding conventions or style guides?","answer":"Yes. You can create a CONVENTIONS.md or .aider.conf.yml file in your project to provide persistent instructions about coding style, preferred libraries, naming conventions, and other guidelines that Aider will follow in every session."}]},"supportStats":{"totalFeatures":13,"supportedCount":4,"partialCount":2,"notSupportedCount":7,"unknownCount":0,"supportPercentage":38}},{"agent":{"id":"claude-code","name":"Claude Code","aliases":["Claude-code","Claude Developer Tool"],"provider":"Anthropic","website":"https://code.claude.com","supported_ide":["Terminal / CLI","VSCode","Cursor","JetBrains","Desktop app","Web","Mobile (iOS/Android)"],"updated_at":"2026-09-30","description":"Claude Code is an agentic coding tool by Anthropic that reads your codebase, edits files, runs commands, and integrates with your development tools. It is available in the terminal, VS Code (and Cursor), JetBrains IDEs, a desktop app, the web, and the Claude mobile app. It defaults to Claude Opus 5.5 on most plans, can switch to Sonnet 5.5, Fable 5.1, or Haiku, and supports a native 1M token context window on Opus 5.5, Sonnet 5.5, and Fable. It can spawn multiple sub-agents in parallel and run scheduled cloud routines.","faq":[{"question":"How does Claude Code integrate with my development environment?","answer":"Claude Code runs in your terminal and also offers a VS Code extension (installable in Cursor too), a JetBrains plugin (IntelliJ IDEA, PyCharm, WebStorm, and others), a desktop app for macOS and Windows, and a web version at claude.ai/code. All surfaces share the same engine, so your CLAUDE.md files, settings, and MCP servers work across them."},{"question":"What programming languages does Claude Code support?","answer":"Claude Code supports a wide range of programming languages, including but not limited to Python, Java, PHP, Rust, and JavaScript frameworks like React and Vue."},{"question":"How do I install Claude Code?","answer":"The recommended method is the native installer: run `curl -fsSL https://claude.ai/install.sh | bash` on macOS, Linux, or WSL, or `irm https://claude.ai/install.ps1 | iex` in Windows PowerShell. It is also available via Homebrew (`brew install --cask claude-code`), WinGet (`winget install Anthropic.ClaudeCode`), and apt, dnf, or apk. Then run `claude` in your project directory."},{"question":"Can Claude Code work with my existing codebase?","answer":"Yes, Claude Code is designed to understand and work with your existing codebase, whether it's a new project or an established one."},{"question":"Does Claude Code require an API key?","answer":"Not necessarily. Most surfaces require either a Claude subscription or an Anthropic Console account; you log in on first use. If the ANTHROPIC_API_KEY environment variable is set, Claude Code asks you to approve that key instead. The terminal CLI, VS Code, and JetBrains also support third-party providers."},{"question":"How does Claude Code handle security and privacy?","answer":"Claude Code prioritizes security by connecting directly to Anthropic’s API without intermediate servers, operating within your terminal, and maintaining awareness of your project structure. It performs actions like editing files and creating commits directly in your environment. Anthropic implements safeguards such as limited retention periods for sensitive information and restricted access to user session data."},{"question":"What are the usage limits for Claude Code?","answer":"Claude Code is included in all paid Claude plans and shares usage limits with Claude chat: Pro ($20/month, or $17/month billed annually), Max (from $100/month), Team Standard ($25/seat/month, or $20 billed annually), Team Premium ($125/seat/month, or $100 billed annually), and Enterprise ($20/seat/month plus usage at API rates). It is not included in the Free plan. See https://claude.com/pricing for details."},{"question":"Can I use Claude Code offline?","answer":"Claude Code requires an internet connection to function, as it communicates with Anthropic's API. There are no offline capabilities mentioned in the current documentation."},{"question":"How does Claude Code compare to other AI coding assistants like GitHub Copilot?","answer":"Claude Code is highly agentic, performing complex tasks like editing files, running tests, and managing git workflows through natural language commands. GitHub Copilot also offers an Agent Mode for multi-file tasks, but it is primarily an extension inside existing IDEs, while Claude Code is terminal-first and also available in IDEs, a desktop app, and the web."},{"question":"What are some best practices for using Claude Code effectively?","answer":"Best practices include providing clear and specific instructions, using the /think command for planning complex tasks, utilizing the memory feature with #tags for persistent context, and regularly reviewing and approving changes made by Claude Code to ensure accuracy."},{"question":"How can I provide feedback or report issues with Claude Code?","answer":"You can provide feedback or report issues using the /bug command within Claude Code or by filing an issue on the official GitHub repository at https://github.com/anthropics/claude-code."},{"question":"Is Claude Code available for enterprise use?","answer":"Yes. Claude Code is included in the Team and Enterprise plans and can also run through third-party providers such as Amazon Bedrock, Google Cloud's Agent Platform, Microsoft Foundry, and the Claude Platform on AWS."},{"question":"Does Claude Code support version control systems like Git?","answer":"Yes, Claude Code supports Git, allowing it to search through git history, resolve merge conflicts, and create commits and pull requests."},{"question":"Can Claude Code generate documentation for my code?","answer":"While Claude Code can provide explanations and answer questions about your code, its primary focus is on coding tasks and workflows. For comprehensive documentation, you may need to use specific commands or integrate with other tools."},{"question":"How does Claude Code handle large codebases?","answer":"Claude Code is designed to understand and work with entire codebases. Opus 5.5, Sonnet 5.5, and Fable support a native 1M token context window; models without 1M support use 200K tokens. For large tasks, it can spawn multiple sub-agents that work on different parts in parallel while a lead agent coordinates and merges the results."},{"question":"Which AI models does Claude Code use?","answer":"On the Anthropic API and on the Pro, Max, Team, and Enterprise plans, Claude Code defaults to Claude Opus 5.5. You can switch models with /model using aliases: opus (Opus 5.5), sonnet (Sonnet 5.5), fable (Fable 5.1), haiku (the fast Haiku model), best (Fable where available), and opusplan (Opus for planning, Sonnet for execution). Model mappings differ on Bedrock, Google Cloud, and Microsoft Foundry."}]},"supportStats":{"totalFeatures":13,"supportedCount":8,"partialCount":3,"notSupportedCount":2,"unknownCount":0,"supportPercentage":73}},{"agent":{"id":"cline","name":"Cline","aliases":["Cline Bot","Cline CLI"],"provider":"Cline Bot Inc. (open source)","website":"https://cline.bot","supported_ide":["VSCode","Cursor","Devin Desktop (formerly Windsurf)","JetBrains","Antigravity","Zed","Neovim","Terminal / CLI","Desktop app"],"updated_at":"2026-09-30","description":"Cline is an open-source (Apache-2.0) autonomous coding agent available in VS Code and compatible editors, JetBrains IDEs, Zed, Neovim, a CLI, and a desktop app for macOS and Windows. It reads and writes files, runs terminal commands, and can use a browser, with every action requiring your explicit approval. It separates planning from execution with Plan and Act modes, supports .clinerules (as well as .cursorrules, .windsurfrules, and AGENTS.md), and connects to MCP servers. You bring your own API key for providers such as Anthropic, OpenAI, Google Gemini, and AWS Bedrock, or use local models via Ollama and LM Studio.","faq":[{"question":"Is Cline free?","answer":"Cline is open source under the Apache 2.0 license. There are three ways to access models: a Cline account with usage-based billing and free model options, ClinePass (a flat $9.99/month subscription with 2-5x the usage on popular open coding models compared to standard API rates), or your own provider credentials (Anthropic, OpenAI, Google Gemini, AWS Bedrock, OpenRouter, and others, or local models via Ollama and LM Studio). Cline can also be configured with a Claude Code subscription."},{"question":"What are Plan and Act modes?","answer":"In Plan mode, Cline can read your codebase, run searches, and discuss strategy, but cannot modify files or execute commands. In Act mode, it keeps the full context from planning and can modify files, run commands, and execute the strategy."},{"question":"Does Cline work with rules from other tools?","answer":"Yes. Besides .clinerules/ (or .cline/rules/), Cline recognizes .cursorrules, .windsurfrules, and AGENTS.md, and shows all detected rules together in its Rules panel."}]},"supportStats":{"totalFeatures":13,"supportedCount":10,"partialCount":1,"notSupportedCount":0,"unknownCount":2,"supportPercentage":81}},{"agent":{"id":"cursor","name":"Cursor","aliases":["Cursor.sh","Cursor AI","Cursor Editor"],"provider":"Cursor.sh","website":"https://cursor.com","supported_ide":["VSCode (fork)","JetBrains (via ACP)"],"updated_at":"2026-09-30","description":"A VSCode-based AI agent for editing, refactoring, and task-based coding with powerful autocomplete and chat features. Cursor 2.0 added an integrated browser with live preview and Chrome DevTools; Cursor 2.2 added a Visual Editor for drag-and-drop UI manipulation. Also available in JetBrains IDEs via Agent Client Protocol.","faq":[{"question":"What models does Cursor support?","answer":"Cursor offers two usage pools. Cursor Models (with more included usage): Composer 2.5 and Grok 4.7 (including a 500k-context variant), Grok 4.6, and Grok 4.5, each with a Fast variant. Other Models, billed at API rates on Pro and higher: Anthropic (Claude Opus 5.5, 5, 4.8, 4.7, 4.6, and 4.5; Sonnet 5.5, 5, 4.6, 4.5, and 4; Haiku 4.5; Fable 5.1 and 5), OpenAI (GPT-5.6 Sol, Terra, and Luna; GPT-5.5; GPT-5.4, Mini, and Nano; and earlier GPT-5 and Codex models), Google (Gemini 3.1 Pro, Gemini 3 Pro, and Gemini Flash models up to 3.8), Z.ai (GLM 5.3 and 5.2), Moonshot (Kimi K3 and Kimi K2.7 Code), and Meta (Muse Spark 1.3). You can switch models per conversation."},{"question":"How does Cursor differ from GitHub Copilot?","answer":"Cursor provides a full IDE experience with integrated chat, multi-file editing, and more advanced AI features. It's a fork of VSCode with AI deeply integrated, while Copilot is an extension that works inside many existing IDEs (VS Code, JetBrains IDEs, Visual Studio, and others) and also offers Agent Mode for autonomous multi-file tasks."},{"question":"Can Cursor work with my existing VSCode extensions?","answer":"Yes, Cursor is compatible with most VSCode extensions since it's built on the VSCode codebase. You can install extensions from the VSCode marketplace just like in regular VSCode."},{"question":"What makes Cursor different from other AI coding assistants like GitHub Copilot or Devin Desktop?","answer":"Cursor is a VS Code-based IDE with deep AI integration, offering natural language interaction, code generation, and multi-file editing. It balances agentic capabilities with a familiar interface, distinguishing it from Copilot's extension-based approach and from Devin Desktop (formerly Windsurf), which centers its IDE on a command center for managing multiple agents."},{"question":"Which AI models does Cursor use?","answer":"Cursor combines its own models (Composer 2.5 and Grok 4.7, 4.6, and 4.5, in the Cursor Models pool) with third-party models from Anthropic (up to Claude Opus 5.5, Sonnet 5.5, and Fable 5.1), OpenAI (up to GPT-5.6), Google (Gemini 3.1 Pro and Gemini Flash up to 3.8), Z.ai, Moonshot, and Meta, allowing users to select models for specific tasks."},{"question":"Is Cursor free to use, and what are the limitations of the free tier?","answer":"Cursor offers a free Hobby plan (no credit card required) with limited Agent requests. Individual plans are Pro ($20/month), Pro Plus ($60/month), and Ultra ($200/month), each with a monthly pool for Cursor Models and another for third-party models charged at API rates. Teams offers Standard seats ($40/user/month) and Premium seats ($120/user/month, with 5x Standard's Agent limits), and adds centralized billing, SAML/OIDC SSO, and Bugbot code reviews; on team plans, third-party model requests carry an extra Cursor Token Rate of $0.25 per million tokens. Enterprise has custom pricing. See https://cursor.com/pricing for details."},{"question":"Can Cursor work with my existing codebases, including private repositories?","answer":"Yes, Cursor can work with existing codebases, including private repositories, as it functions as a full IDE based on VS Code, accessing local or remote folders."},{"question":"Does Cursor support all programming languages?","answer":"Cursor supports all programming languages supported by VS Code, covering a wide range of languages and frameworks."},{"question":"How does Cursor handle security and privacy, especially for private code?","answer":"Cursor offers a Privacy Mode ensuring code is not stored remotely without consent. It is SOC 2 certified, emphasizing robust security practices. See https://cursor.com/en for details."},{"question":"What are the system requirements for running Cursor?","answer":"As a VS Code fork, Cursor's system requirements align with VS Code, running on modern versions of Windows, macOS, or Linux. Check https://code.visualstudio.com/docs/supporting/requirements for VS Code requirements."},{"question":"How can I provide feedback or report issues with Cursor?","answer":"Feedback and issues can be reported via Cursor's GitHub repository at https://github.com/cursor/cursor or through Anysphere's support channels."},{"question":"Does Cursor integrate with version control systems like Git?","answer":"Yes, Cursor supports Git integration, leveraging VS Code's capabilities to manage code repositories within the IDE."},{"question":"Can Cursor help with code reviews or testing?","answer":"Cursor's AI features, like error detection and code suggestions, can assist with code reviews and testing, though specific tools for these tasks may require additional extensions."},{"question":"How does Cursor's pricing compare to other AI coding tools?","answer":"Cursor Pro is $20/month. GitHub Copilot Pro is $10/month (cheapest entry point) and Devin Desktop Pro is $20/month. For power users, Cursor Pro Plus ($60/month) and Ultra ($200/month) include larger usage pools. See https://cursor.com/pricing for the latest."},{"question":"Is there a way to customize Cursor's suggestions or behavior?","answer":"Cursor allows model selection and extensive editor customization, similar to VS Code, enabling tailored suggestion behavior."},{"question":"Can I use Cursor offline?","answer":"Basic editing is possible offline, but AI features require an internet connection to access Anysphere's servers."},{"question":"How does Cursor handle large codebases or complex projects?","answer":"Cursor uses custom retrieval models to understand entire codebases, making it effective for large and complex projects."},{"question":"What kind of support does Cursor offer for users?","answer":"Support is available through Cursor's GitHub repository, official documentation at cursor.com/docs, and Anysphere's support channels, with enhanced options for Pro and Enterprise users."},{"question":"What are Background Agents in Cursor?","answer":"Background Agents (available since Cursor 0.50+) allow you to run multiple AI agent tasks in parallel, each in its own isolated environment. You can kick off a long-running task and continue coding while the agent works independently, then review results when it's done."},{"question":"Does Cursor have a built-in browser or web preview?","answer":"Yes. Cursor 2.0 (November 2025) introduced an integrated browser panel with Chrome DevTools access. You can see your running web application live inside the editor, and the AI agent can directly read console errors and network requests to debug issues."}]},"supportStats":{"totalFeatures":13,"supportedCount":13,"partialCount":0,"notSupportedCount":0,"unknownCount":0,"supportPercentage":100}},{"agent":{"id":"devin-ai","name":"Devin","aliases":["Devin AI","Cognition Devin","Devin software engineer"],"provider":"Cognition AI","website":"https://devin.ai","supported_ide":["Browser","Slack","VS Code (remote)","Devin Desktop"],"updated_at":"2026-09-30","description":"Devin is an autonomous AI software engineer by Cognition AI, marketed as the first AI agent capable of completing end-to-end software engineering tasks independently. Devin works in a fully isolated cloud environment equipped with a shell, code editor, and browser. It can plan and execute complex, multi-step engineering tasks — from cloning a repo and writing code to deploying applications and fixing bugs — with minimal human intervention. Devin is accessible via web, Slack integration, and a VS Code extension, and cloud sessions can be managed from Devin Desktop (formerly Windsurf).","faq":[{"question":"What is Devin and what makes it different from other AI coding tools?","answer":"Devin is described as an autonomous AI software engineer. Unlike pair programmers (Copilot, Cursor) that assist while you code, Devin works independently to complete full engineering tasks: setting up environments, writing and testing code, debugging, and deploying. It operates in a cloud sandbox with its own shell, browser, and editor, and reports back when the task is done."},{"question":"Which AI model powers Devin?","answer":"Devin supports models from several providers: Anthropic (Claude), OpenAI (GPT), Google (Gemini), SpaceXAI, open-source models such as DeepSeek, Kimi, and GLM, and Cognition's own SWE models. The Free plan has limited model availability; Pro and higher unlock full model availability, and Pro includes free use of SWE-2 Free in Devin Desktop and the CLI through October 16, 2026."},{"question":"How much does Devin cost?","answer":"Devin's plans are shared with Devin Desktop: Free ($0, light agent quota and unlimited Tab completions and inline edits), Pro ($20/month, increased quotas, full model availability, and cloud agents), Max ($200/month, significantly higher quotas), Teams ($80/month base plus $40/month per developer seat, up to 200 users), and Enterprise (custom pricing, SSO, and VPC deployment). Costs vary by model and task complexity. See https://devin.ai/pricing for details."},{"question":"What kinds of tasks can Devin handle autonomously?","answer":"Devin can handle: setting up codebases from scratch, implementing features end-to-end, fixing bugs from issue descriptions, writing and running tests, deploying to cloud providers, refactoring code, updating dependencies, writing documentation, and completing tasks from GitHub issues or Jira tickets."},{"question":"How does Devin interact with my codebase?","answer":"Devin clones your repository into a private cloud environment and has access to a terminal, code editor, and browser. It can run commands, install packages, write code, run tests, and browse documentation. You observe its progress in real time via a session dashboard and can intervene or provide feedback at any point."},{"question":"Can I use Devin with Slack?","answer":"Yes. Devin integrates with Slack, allowing team members to assign tasks to Devin directly from Slack messages. Devin reports progress and asks clarifying questions in the Slack thread."},{"question":"Does Devin support GitHub and Jira integration?","answer":"Yes. Devin integrates with GitHub to work on issues and pull requests directly. It can also connect to Jira to pick up tickets and implement the requested changes. These integrations allow Devin to work asynchronously on your backlog."},{"question":"Is Devin's cloud environment secure?","answer":"Devin runs in isolated cloud environments (one per session) with no cross-session data sharing. Cognition AI uses encrypted credential storage and does not retain repository code after sessions end. Enterprise plans include additional compliance controls."},{"question":"Can Devin browse the web?","answer":"Yes. Devin has access to a cloud browser in its sandbox environment. It uses this to read documentation, search for solutions, look up API references, and navigate to any web resources needed to complete a task."},{"question":"What are Devin's limitations?","answer":"Devin performs best on self-contained tasks with clear requirements. It can struggle with deeply ambiguous tasks requiring business context, tasks requiring persistent human decisions mid-stream, proprietary internal tools it has no documentation for, and very large tasks beyond 4-8 hours of human effort. Human review of all PRs before merging is strongly recommended."}]},"supportStats":{"totalFeatures":13,"supportedCount":7,"partialCount":4,"notSupportedCount":2,"unknownCount":0,"supportPercentage":69}},{"agent":{"id":"devin-desktop","name":"Devin Desktop","aliases":["Windsurf","Windsurf IDE","Devin IDE"],"provider":"Cognition AI","website":"https://devin.ai/desktop","supported_ide":["Devin Desktop (custom IDE)","JetBrains (Windsurf plugin)"],"updated_at":"2026-09-30","description":"Devin Desktop is the new name for Windsurf, the AI-native IDE from Cognition AI. Built on the Windsurf IDE foundation, it puts an Agent Command Center front and center, with Spaces for sharing context and Git worktrees across agents and a Kanban view of agent sessions, while keeping the full IDE experience, including Tab completions, the Cascade agent, and the Devin Local agent. Local work can be handed off to Devin cloud agents.","faq":[{"question":"Is Devin Desktop the same as Windsurf?","answer":"Yes. Devin Desktop is the new name for Windsurf. Existing installations were updated automatically with no changes to plans, pricing, extensions, or settings, and the Windsurf plugin for JetBrains IDEs (IntelliJ IDEA, PyCharm, WebStorm, and more) remains available for download."},{"question":"How much does Devin Desktop cost?","answer":"Devin Desktop shares its plans with Devin: Free ($0, light agent quota and unlimited Tab completions and inline edits), Pro ($20/month), Max ($200/month), Teams ($80/month base plus $40/month per developer seat), and Enterprise (custom pricing). See https://devin.ai/pricing for details."},{"question":"Which AI models does Devin Desktop use?","answer":"The Devin Desktop model catalog includes Cognition's SWE-2, SWE-1.7, and SWE-1.6 plus the Adaptive router; Anthropic Claude Opus 4.5 to 5.5, Sonnet 4.5, 4.6, and 5, Haiku 4.5, and Fable 5 and 5.1; OpenAI models from GPT-4o and GPT-4.1 up to GPT-5.6 (Sol, Terra, Luna) and GPT-6 (Astra, Sol, Luna); Google Gemini 2.5 Pro, 3.1 Pro, and Flash models up to 3.8; xAI Grok; DeepSeek V4; Moonshot Kimi K2.6, K2.7, and K3; Z.ai GLM-5.2 and 5.3; and NVIDIA Nemotron 3 Ultra. Availability depends on your plan: the Free plan has limited model availability."},{"question":"What agents are included in Devin Desktop?","answer":"Devin Desktop includes Cascade, an agentic assistant with Code and Chat modes, tool calling, checkpoints, and built-in planning, and the Devin Local agent, which brings the Devin CLI harness with subagents, sandboxing, plan mode, worktrees, and permissions. Work can also be handed off to Devin cloud agents."}]},"supportStats":{"totalFeatures":13,"supportedCount":11,"partialCount":2,"notSupportedCount":0,"unknownCount":0,"supportPercentage":92}},{"agent":{"id":"gemini-cli","name":"Gemini CLI","aliases":["Google Gemini CLI","gemini"],"provider":"Google","website":"https://github.com/google-gemini/gemini-cli","supported_ide":["Terminal / CLI","VSCode (Companion extension)","JetBrains (ACP)","Zed (ACP)"],"updated_at":"2026-09-30","description":"Gemini CLI is Google's open-source (Apache-2.0) AI agent that brings Gemini models directly into your terminal. It has built-in tools for file operations, shell commands, web fetch, and Google Search grounding, supports MCP servers and GEMINI.md context files, and uses Gemini 3 models with a 1M token context window. It integrates with VS Code-compatible editors through the Gemini CLI Companion extension and with JetBrains IDEs and Zed through the ACP Agent Registry.","faq":[{"question":"Is Gemini CLI free?","answer":"Yes, within quotas: signing in with a Google account gives 60 requests per minute and 1,000 requests per day, and a Gemini API key gives 1,000 requests per day. Paid usage-based billing and Vertex AI options offer higher limits."},{"question":"How do I install Gemini CLI?","answer":"Run it without installing via npx @google/gemini-cli, or install it globally with npm install -g @google/gemini-cli. It is also available via Homebrew, MacPorts, and Anaconda."},{"question":"What is GEMINI.md?","answer":"GEMINI.md files provide instructional context. Gemini CLI loads a global ~/.gemini/GEMINI.md, GEMINI.md files in your workspace directories and their parents, and, just in time, GEMINI.md files in directories that tools access. Use /memory show to see the loaded context."}]},"supportStats":{"totalFeatures":13,"supportedCount":8,"partialCount":2,"notSupportedCount":0,"unknownCount":3,"supportPercentage":69}},{"agent":{"id":"github-copilot","name":"GitHub Copilot","aliases":["Copilot","GitHub Copilot X","Copilot Chat"],"provider":"GitHub","website":"https://github.com/features/copilot","supported_ide":["VSCode","JetBrains IDEs","Neovim","Emacs","Visual Studio"],"description":"An AI-powered coding assistant deeply integrated with GitHub. Evolved from a code completion tool into a full coding agent with Agent Mode (GA 2025), MCP support, custom instruction files, and multi-model support across Claude, GPT, Gemini, Grok, Kimi, and MAI-Code models.","updated_at":"2026-09-30","faq":[{"question":"What models does GitHub Copilot use?","answer":"Paid plans can choose from Claude Haiku 4.5, Claude Opus 4.7, 4.8, 5, and 5.5, Claude Sonnet 4.6, 5, and 5.5, Claude Fable 5 and 5.1, Gemini 3.5, 3.6, 3.7, and 3.8 Flash, GPT-5 mini, GPT-5.3-Codex, GPT-5.4 (and mini and nano), GPT-5.5, GPT-5.6 Luna, Sol, and Terra, GPT-6 Astra, Luna, and Sol, GPT-6.1 Sol, Grok 4.5, 4.6, and 4.7, Kimi K2.7 Code, Kimi K3, and MAI-Code-1.1-Flash. Pro includes a selection of models, Pro+ and Business include premium models, and Max and Enterprise get priority access to premium models. Copilot Free and Student use auto model selection only."},{"question":"How does GitHub Copilot differ from other AI coding assistants?","answer":"GitHub Copilot has evolved significantly. It now includes Agent Mode (autonomous multi-file coding), MCP server support, and custom instruction files (.github/copilot-instructions.md). It remains the most widely deployed AI coding tool with the broadest IDE support."},{"question":"Can GitHub Copilot work with private repositories?","answer":"Yes, GitHub Copilot can work with private repositories, but it doesn't train on private code. It uses the context of your current file and repository to provide better suggestions while respecting privacy."},{"question":"How does GitHub Copilot learn from my code?","answer":"GitHub Copilot does not learn from your private code in terms of training its model. It is trained on public code repositories. However, it uses the context of your current file and repository to provide better suggestions, adapting to your coding style without storing your private code."},{"question":"Can I customize the suggestions in GitHub Copilot?","answer":"Yes. You can create a .github/copilot-instructions.md file in your repository to provide persistent, natural-language instructions (coding standards, preferred libraries, response style). Since November 2025, multiple .instructions.md files with YAML frontmatter allow path-specific and agent-specific rules."},{"question":"Does GitHub Copilot support all programming languages?","answer":"GitHub Copilot supports many programming languages, particularly popular ones like JavaScript, Python, TypeScript, Ruby, C++, and Rust. The quality of suggestions varies, with better support for languages with more public code available for training."},{"question":"How does GitHub Copilot handle security and privacy?","answer":"GitHub Copilot is designed with privacy in mind, but under regulations like GDPR, it may output synthesized personal data. Administrators can exclude files from Copilot access, and users can block suggestions matching public code to enhance security."},{"question":"Is GitHub Copilot available for individual developers or only for teams?","answer":"GitHub Copilot is available for both individual developers and businesses. A free tier with limited functionality is offered for individuals, alongside paid plans like Copilot Pro and Enterprise for additional features."},{"question":"What are the system requirements for using GitHub Copilot?","answer":"GitHub Copilot operates as an extension for IDEs like Visual Studio Code, Visual Studio, and JetBrains IDEs. Its system requirements match those of the supported IDEs, which typically run on modern versions of Windows, macOS, or Linux."},{"question":"How can I provide feedback to improve GitHub Copilot?","answer":"Feedback can be provided via the 'Send Copilot Completion Feedback' option when hovering over a completion, the Feedback action in next edit suggestions, or the Issue Reporter in VS Code. Visit https://github.com/microsoft/vscode for more details."},{"question":"Does GitHub Copilot work with version control systems like Git?","answer":"Yes, GitHub Copilot integrates seamlessly with Git repositories through its deep integration with GitHub, providing context-aware suggestions based on your repository's structure."},{"question":"Can GitHub Copilot help with code reviews?","answer":"Yes, Copilot Code Review provides AI-powered feedback on pull requests, supporting all programming languages in public preview as of May 2025. See https://github.blog/changelog/2025-05-07-copilot-code-review-now-supports-all-languages-in-public-preview/."},{"question":"How does GitHub Copilot compare to other code completion tools like Tabnine or Amazon CodeWhisperer?","answer":"GitHub Copilot's strengths are its deep GitHub integration, broadest IDE support (VS Code, JetBrains, Neovim, Emacs, Visual Studio), multi-model selection, and Agent Mode. Amazon Q Developer (successor to CodeWhisperer) is being retired: its IDE plugins reach end of support on April 30, 2027, and AWS directs users to Kiro. Tabnine specializes in on-premise enterprise deployments with zero data transmission."},{"question":"Is there a way to disable GitHub Copilot temporarily?","answer":"Yes, you can disable GitHub Copilot via the Copilot menu in your IDE's title bar or by toggling settings in the IDE, such as in VS Code's status dashboard."},{"question":"Does GitHub Copilot support keyboard shortcuts?","answer":"Yes, GitHub Copilot offers default keyboard shortcuts for accepting suggestions, which can be customized in supported IDEs like VS Code or JetBrains."},{"question":"Can I use GitHub Copilot offline?","answer":"No, GitHub Copilot requires an internet connection to communicate with GitHub's servers for generating suggestions."},{"question":"How does GitHub Copilot handle large codebases?","answer":"GitHub Copilot uses the context of your current file and repository to provide suggestions for large codebases, though performance may vary based on codebase complexity."},{"question":"What is the pricing model for GitHub Copilot?","answer":"GitHub Copilot usage is measured in monthly GitHub AI Credits. Individual plans: Free (an allowance of AI credits and up to 2,000 completions per month), Student (free for verified students), Pro ($10/month, 1,500 credits; free for verified teachers and maintainers of popular open source projects), Pro+ ($39/month, 7,000 credits, premium models), and Max ($100/month, 20,000 credits, priority access to premium models). Organization plans: Business ($19/seat/month, 1,900 credits per user) and Enterprise ($39/seat/month, 3,900 credits per user), with extra usage at $0.01 per credit. Code completions and next edit suggestions are not billed in AI credits on paid plans. See https://docs.github.com/en/copilot/get-started/plans for details."},{"question":"What is Agent Mode in GitHub Copilot?","answer":"Agent Mode (generally available in VS Code since February 2025) allows Copilot to autonomously complete multi-step tasks: it reads your codebase, creates a plan, writes and modifies files, runs commands, and iterates until the task is done — all with your approval at each step."},{"question":"Does GitHub Copilot support MCP servers?","answer":"Yes. GitHub Copilot supports MCP (Model Context Protocol) servers in VS Code, JetBrains, and Xcode as of 2025. MCP tools are accessible directly from the Agent Mode panel, allowing Copilot to interact with databases, APIs, and custom tools."}]},"supportStats":{"totalFeatures":13,"supportedCount":7,"partialCount":3,"notSupportedCount":3,"unknownCount":0,"supportPercentage":65}},{"agent":{"id":"google-antigravity","name":"Google Antigravity","aliases":["Antigravity","Antigravity IDE","Google Antigravity 2.0"],"provider":"Google","website":"https://antigravity.google","supported_ide":["Antigravity (desktop IDE)","Terminal / CLI","VSCode","Visual Studio","JetBrains","Zed","Xcode"],"updated_at":"2026-09-30","description":"Google Antigravity is Google's agentic development platform, announced on November 20, 2025. It combines an AI-powered IDE with an agent-first Manager surface where you can spawn, orchestrate, and observe multiple agents working asynchronously across workspaces. Agents can autonomously plan and execute tasks across the editor, terminal, and browser, and report progress through artifacts such as implementation plans, screenshots, and walkthroughs. It is also available as a CLI, an SDK, and extensions for VS Code, Visual Studio, JetBrains, Zed, and Xcode.","faq":[{"question":"How much does Google Antigravity cost?","answer":"Antigravity is generally available. The Individual plan is free and includes agent models, unlimited Tab completions and command requests, and basic weekly rate limits. Google AI Pro adds more generous rate limits and a flexible AI credit pool, and Google AI Ultra adds higher access to the latest Gemini models. Organizations can use it through Google Cloud with consumption-based pricing. See https://antigravity.google/pricing for details."},{"question":"Which AI models does Antigravity support?","answer":"The free Individual plan includes Gemini 3.8 Flash, Gemini 3.7 Flash, Gemini 3.6 Flash, Gemini 3.1 Pro, Claude Sonnet 4.6, Claude Opus 4.6, and gpt-oss-120b as agent models. Google AI Ultra adds higher access to the latest Gemini models."},{"question":"Which operating systems does Antigravity support?","answer":"Antigravity is cross-platform and runs on macOS, Windows, and Linux."},{"question":"What is the Agent Manager?","answer":"The Manager surface is a dedicated interface where you can spawn, orchestrate, and observe multiple agents working asynchronously across different workspaces."}]},"supportStats":{"totalFeatures":13,"supportedCount":9,"partialCount":2,"notSupportedCount":0,"unknownCount":2,"supportPercentage":77}},{"agent":{"id":"google-jules","name":"Google Jules","aliases":["Jules","Jules by Google","Google Jules agent"],"provider":"Google","website":"https://jules.google","supported_ide":["GitHub (web)","Browser","Jules Tools (CLI)","Jules API"],"updated_at":"2026-09-30","description":"Google Jules is an asynchronous AI coding agent by Google, powered by Gemini 3 Flash (all tiers) and Gemini 3.1 Pro (Google AI Pro plan). Unlike real-time pair programmers, Jules works autonomously in the background on GitHub issues and pull requests. You assign it a task (bug fix, feature, refactor), Jules clones your repo into a sandboxed cloud environment, implements the changes, runs tests, and opens a pull request for your review. Jules has been generally available since August 2025 and also offers a command-line tool (Jules Tools), a REST API, scheduled tasks, automatic CI fixes, and MCP integrations.","faq":[{"question":"What is Google Jules and how does it work?","answer":"Jules is an asynchronous AI coding agent that works on GitHub issues and tasks in the background. You assign Jules a task — such as fixing a bug or implementing a feature — and it independently clones your repository into an isolated cloud environment, writes the code, runs tests, and opens a pull request for your review. You review and merge the PR when satisfied."},{"question":"Which AI model powers Google Jules?","answer":"Since January 30, 2026, Gemini 3 Flash is the base model in Jules for all users on all tiers. Since March 9, 2026, Gemini 3.1 Pro is available as the default model for complex problem-solving for Google AI Pro plan users."},{"question":"Is Google Jules free to use?","answer":"Jules has been generally available since August 6, 2025 and can be used with a Google account at jules.google. The free plan allows 15 tasks per rolling 24 hours and 3 concurrent tasks; Jules in Pro (included with Google AI Pro) allows 100 daily and 15 concurrent tasks; Jules in Ultra (included with Google AI Ultra) allows 300 daily and 60 concurrent tasks. See https://jules.google/docs/usage-limits/ for details."},{"question":"Does Jules work with any GitHub repository?","answer":"Jules integrates with GitHub and can work with any public or private GitHub repository you grant it access to. It clones your repo into a sandboxed cloud environment for each task, so it does not run on your local machine."},{"question":"How is Jules different from GitHub Copilot or Cursor?","answer":"Jules is fundamentally asynchronous — you do not interact with it in real time. You assign a task and come back later to review a pull request. Tools like Copilot and Cursor work synchronously in your IDE as you type. Jules is better suited for delegating entire self-contained tasks rather than pair programming."},{"question":"What kinds of tasks is Jules best at?","answer":"Jules excels at self-contained, well-defined tasks: fixing bugs from issue descriptions, implementing small features, writing or updating unit tests, updating dependencies, and performing code migrations. Tasks requiring interactive decisions or ambiguous requirements are better handled by synchronous agents."},{"question":"Does Jules run tests before opening a pull request?","answer":"Yes. Jules runs the repository's existing test suite as part of its task execution in the sandboxed environment. It uses test results to validate its implementation and may iterate on the code if tests fail."},{"question":"Can Jules access external APIs or databases?","answer":"Jules operates in a sandboxed cloud environment and does not have access to external services, live databases, or production systems. It works only with the code and tests in your repository."},{"question":"Is Jules available outside the United States?","answer":"Jules has been generally available since August 6, 2025. Check jules.google for the latest access information in your region."},{"question":"Does Jules support languages other than English?","answer":"Jules understands task descriptions written in multiple languages, as the underlying Gemini models are multilingual. However, the Jules interface and documentation are primarily in English."}]},"supportStats":{"totalFeatures":13,"supportedCount":5,"partialCount":1,"notSupportedCount":6,"unknownCount":1,"supportPercentage":42}},{"agent":{"id":"kiro","name":"Kiro","aliases":["Kiro IDE","Kiro CLI","AWS Kiro"],"provider":"Amazon Web Services","website":"https://kiro.dev","supported_ide":["Kiro IDE","Terminal / CLI","Web"],"updated_at":"2026-09-30","description":"Kiro is an agentic development environment from AWS, available as an IDE, a CLI, and on the web, built for spec-driven development: it turns a feature request into requirements, a technical design, and implementation tasks before writing code. It supports steering files (including AGENTS.md), agent hooks, MCP servers, and an Autopilot mode that can create and modify files autonomously. AWS recommends Kiro as the successor to Amazon Q Developer.","faq":[{"question":"How much does Kiro cost?","answer":"Kiro offers Free ($0/month, 50 credits), Pro ($20/user/month, 1,000 credits), Pro+ ($40/user/month, 2,000 credits), Pro Max ($100/user/month, 5,000 credits), and Power ($200/user/month, 10,000 credits). Paid tiers can buy extra credits at $0.04 per credit. Credits are consumed across the IDE, CLI, and web. See https://kiro.dev/pricing/ for details."},{"question":"Which AI models does Kiro support?","answer":"Paid plans (Pro, Pro+, Pro Max, and Power) include Auto (an agent that uses a mix of frontier models), Claude Sonnet 5, Claude Opus 5, Claude Opus 4.8, Claude Sonnet 4.6, Claude Haiku 4.5, GPT-5.6 (Sol, Terra, and Luna, with a 1M token context window), and open-weight models such as DeepSeek 3.2, MiniMax M2.5, GLM-5, and Qwen3 Coder Next. Claude Opus 5.5 has been rolling out to paid plans since September 28, 2026 as an experimental model in selected AWS regions. Claude Fable 5.1 is available only to enterprise customers with Model Governance enabled. The Free plan includes Claude Sonnet 4.5 and open-weight models such as Qwen3 Coder Next, DeepSeek 3.2, and MiniMax M2.1."},{"question":"What is spec-driven development in Kiro?","answer":"Specs are structured artifacts that formalize the development of features and bug fixes. Kiro works through requirements (requirements.md), a technical design (design.md), and a list of implementation tasks (tasks.md) with built-in progress tracking."},{"question":"How does Kiro relate to Amazon Q Developer?","answer":"AWS announced the end of support for the Amazon Q Developer IDE plugins and paid subscriptions: new signups were blocked on May 15, 2026, and support ends on April 30, 2027. AWS advises customers to transition to Kiro during that period."}]},"supportStats":{"totalFeatures":13,"supportedCount":10,"partialCount":0,"notSupportedCount":0,"unknownCount":3,"supportPercentage":77}},{"agent":{"id":"openai-codex","name":"OpenAI Codex","aliases":["Codex","Codex CLI","Codex IDE extension"],"provider":"OpenAI","website":"https://chatgpt.com/codex","supported_ide":["Terminal / CLI","VSCode","Cursor","Devin Desktop (formerly Windsurf)","Desktop app","Web"],"updated_at":"2026-09-30","description":"OpenAI Codex is OpenAI's coding agent. The open-source Codex CLI (Apache-2.0) runs locally in your terminal on macOS, Linux, and Windows, and Codex is also available as an IDE extension (VS Code, Cursor, and Devin Desktop (formerly Windsurf)), in the desktop app, and in the cloud at chatgpt.com/codex, with GitHub and Slack integrations. By default it can read files, make edits, and run commands inside a sandboxed workspace, asking for approval before editing outside it or accessing the network.","faq":[{"question":"Which plans include Codex?","answer":"Codex is included in the ChatGPT Plus, Pro, Business, Edu, and Enterprise plans. Free and Go plans can also use Codex in the ChatGPT desktop app with GPT-6 Luna, where available."},{"question":"How do I install the Codex CLI?","answer":"Install it with npm (npm install -g @openai/codex), Homebrew (brew install --cask codex), the install scripts from chatgpt.com, or a binary download from GitHub. It runs on macOS (Apple Silicon and x86_64), Linux (x86_64 and arm64), and Windows."},{"question":"Which models does Codex use?","answer":"Since late September 2026, OpenAI has been rolling out GPT-6 Sol and GPT-6 Luna in Codex, at lower token prices than their GPT-5.6 predecessors. For complex coding and professional work, OpenAI recommends GPT-6.1 Sol. The documentation does not name a single default model."},{"question":"Is Codex safe to run on my machine?","answer":"By default Codex uses the workspace-write sandbox: it can read files, make edits, and run commands in the active workspace, but needs approval to edit outside it or to access the network, and network access is off by default. A read-only mode is also available."}]},"supportStats":{"totalFeatures":13,"supportedCount":9,"partialCount":2,"notSupportedCount":0,"unknownCount":2,"supportPercentage":77}},{"agent":{"id":"opencode","name":"OpenCode","aliases":["opencode","OpenCode AI"],"provider":"anomalyco (open source)","website":"https://opencode.ai","supported_ide":["Terminal (TUI)","Desktop app","Web","VSCode","Cursor","Devin Desktop (formerly Windsurf)","VSCodium"],"updated_at":"2026-09-30","description":"OpenCode is an open-source (MIT) AI coding agent with a terminal UI, a desktop app for macOS, Windows, and Linux, a web interface, and IDE integration for VS Code and its forks. It works with any LLM provider through your own API keys and offers two built-in agents: Build, with full access, and Plan, a read-only mode that suggests how to implement a change. It supports AGENTS.md rules, MCP servers, and LSP servers, and is one of the most-starred coding agents on GitHub, with more than 200,000 stars.","faq":[{"question":"Is OpenCode free?","answer":"OpenCode itself is open source under the MIT license. You connect it to any LLM provider by configuring their API keys, and OpenCode Zen offers a curated list of recommended models."},{"question":"What is the difference between Build and Plan?","answer":"Build is the default agent with full access. Plan mode disables its ability to make changes and instead suggests how it will implement the feature; switch between them with the Tab key."},{"question":"Where does OpenCode read project instructions from?","answer":"Run /init to create an AGENTS.md in the project root. Global rules go in ~/.config/opencode/AGENTS.md, and the instructions option in opencode.json can include additional rule files using glob patterns."}]},"supportStats":{"totalFeatures":13,"supportedCount":7,"partialCount":3,"notSupportedCount":0,"unknownCount":3,"supportPercentage":65}},{"agent":{"id":"replit-agent","name":"Replit Agent","aliases":["Replit AI Agent","Replit Ghostwriter Agent","Replit AI"],"provider":"Replit","website":"https://replit.com/ai","supported_ide":["Replit (browser)","Replit Desktop"],"updated_at":"2026-09-30","description":"Replit Agent is an AI coding agent built into the Replit cloud development environment. It can build, run, and deploy full applications entirely from natural language prompts, with no local setup required. Replit Agent handles the entire development lifecycle — from writing code and installing dependencies to configuring databases and deploying to production — all within Replit's browser-based IDE. It is designed to make software development accessible to non-developers and to accelerate prototyping for professionals.","faq":[{"question":"What is Replit Agent and what can it build?","answer":"Replit Agent is an AI that builds full applications from a text prompt inside Replit's cloud IDE. You describe what you want — a web app, REST API, Discord bot, database-backed tool, etc. — and Replit Agent writes the code, installs packages, sets up databases, and deploys the app. No local environment setup is required."},{"question":"Which AI models power Replit Agent?","answer":"Replit Agent is powered by large language models from multiple providers. Since September 2026, users can choose GPT-6 Sol, GPT-6 Luna Fast, and Claude Opus 5.5 when building with Agent; availability depends on your plan and workspace policies."},{"question":"Do I need coding experience to use Replit Agent?","answer":"No. Replit Agent is designed to be usable by non-developers. You describe what you want in plain English and the agent handles the implementation. However, developers can also use it to accelerate prototyping by iterating on the generated code directly in Replit's editor."},{"question":"How does Replit Agent deploy applications?","answer":"Replit Agent can deploy applications directly to Replit's cloud hosting infrastructure (Replit Deployments) with a single click or as part of the agent workflow. Deployed apps get a public URL instantly. Replit supports web apps, APIs, bots, and scheduled jobs."},{"question":"Does Replit Agent support databases?","answer":"Yes. Replit Agent can provision and configure Replit's built-in database options: Replit DB (a simple key-value store), PostgreSQL (via Neon), and other database types. It writes the connection code and schema setup automatically."},{"question":"How much does Replit Agent cost?","answer":"Replit's paid plans are Core ($20/month, or $18/month billed annually, including $20 of usage toward powerful models), Pro ($100/month, or $90/month billed annually, including $100 of usage, 10 parallel agents, and up to 15 collaborators), and Enterprise (custom pricing with SSO/SAML and single-tenant environments). See https://replit.com/pricing for details."},{"question":"Can Replit Agent work with an existing codebase?","answer":"Yes. You can import a GitHub repository into Replit and then use Replit Agent to modify, extend, or debug the existing code. Replit Agent reads the imported codebase and can make targeted edits or implement new features on top of it."},{"question":"Does Replit Agent support version control?","answer":"Replit has built-in Git integration. Replit Agent can commit changes to git history within Replit's environment. You can also push changes to GitHub from Replit. Full two-way GitHub sync is available."},{"question":"What programming languages does Replit Agent support?","answer":"Replit Agent supports any language available on Replit's platform, including Python, JavaScript/TypeScript, Node.js, Ruby, Go, Java, C/C++, Rust, PHP, and many others. Replit runs any language in its cloud environment without local installation."},{"question":"How does Replit Agent compare to tools like Devin or Bolt?","answer":"Replit Agent is most similar to Bolt.new and v0 in that it targets rapid app creation from prompts in a browser-based environment with instant deployment. Devin is more focused on complex, long-horizon engineering tasks on existing codebases. Replit Agent's unique advantage is Replit's fully integrated cloud IDE, built-in hosting, and database — all in one platform without needing any external services."}]},"supportStats":{"totalFeatures":13,"supportedCount":8,"partialCount":1,"notSupportedCount":4,"unknownCount":0,"supportPercentage":65}},{"agent":{"id":"zed-ai","name":"Zed AI","aliases":["Zed","Zed Editor AI","Zed Assistant"],"provider":"Zed Industries","website":"https://zed.dev","supported_ide":["Zed"],"updated_at":"2026-09-30","description":"Zed AI is the AI assistant built natively into the Zed code editor, a high-performance, GPU-accelerated collaborative editor written in Rust. Zed AI supports inline editing, an AI chat panel, multi-model selection (Zed-hosted models or your own API keys from Anthropic, OpenAI, Google, and others), external agents such as Claude Agent and Codex CLI, and an agentic mode that can autonomously edit files, run terminal commands, and use MCP tools. Zed is designed for speed and collaboration, with multiplayer editing capabilities.","faq":[{"question":"What is Zed AI and how does it differ from other AI coding tools?","answer":"Zed AI is the AI layer built directly into the Zed code editor — it is not a plugin or extension but a first-class feature of the editor itself. Zed is built in Rust for maximum performance, and its AI features include inline editing, an AI assistant panel, and an agentic mode. Unlike tools like Cursor (which is VS Code-based) or Claude Code (which is terminal-based), Zed AI runs inside a native GPU-accelerated editor."},{"question":"Which AI models does Zed AI support?","answer":"Zed offers Zed-hosted AI models (included with Pro credits and billed by usage beyond that) and lets you bring your own API keys from Amazon Bedrock, Anthropic, GitHub Copilot, DeepSeek, Google AI, LM Studio, Mistral, Ollama, OpenAI, OpenRouter, Vercel, and more. You can also run external agents such as Claude Agent and Codex CLI inside Zed."},{"question":"Is Zed AI free to use?","answer":"Zed's Personal plan is free forever and includes 2,000 accepted edit predictions, plus unlimited use with your own API keys or external agents. Pro costs $10/month (14-day free trial) and includes unlimited edit predictions and $5 of monthly token credits for Zed-hosted models, with usage beyond that billed at API list price plus 10%. Business costs $30/seat/month and adds org-wide AI model policies and data governance. See https://zed.dev/pricing for details."},{"question":"What is Zed's agentic mode?","answer":"Zed's agentic mode (available from Zed 0.170+) allows the AI assistant to autonomously edit files, run terminal commands, and use MCP-connected tools to complete multi-step coding tasks. It operates within the editor with user approval at key steps."},{"question":"Does Zed AI support MCP (Model Context Protocol)?","answer":"Yes. Zed supports MCP servers, allowing the AI assistant to connect to external tools such as databases, APIs, documentation servers, and custom tooling. MCP tools appear as callable functions in the assistant panel and agentic mode."},{"question":"Does Zed support multiplayer or collaborative coding?","answer":"Yes. Zed was built from the ground up for real-time multiplayer editing — multiple developers can edit the same file simultaneously, similar to Google Docs. This collaboration layer is separate from Zed AI but can be combined with it."},{"question":"Can I use Zed AI with local models?","answer":"Yes. Zed integrates with Ollama, allowing you to run local models (like Llama 3, Mistral, Phi, etc.) entirely on your machine with no API costs or data transmission. Configure Ollama as a provider in Zed's settings."},{"question":"Which operating systems does Zed support?","answer":"Zed supports macOS (native) and Linux. A Windows version is in development as of 2026. Zed is not available as a web app or browser extension."},{"question":"How does Zed AI compare to Cursor?","answer":"Both offer inline AI editing and multi-model support, but their foundations differ. Cursor is built on VS Code and inherits its vast extension ecosystem. Zed is a fully custom, high-performance Rust editor that is much faster but has a smaller extension ecosystem. Zed has native multiplayer; Cursor does not. Cursor has a more mature agentic experience as of early 2026."},{"question":"Does Zed AI have an instruction file for project-specific context?","answer":"Yes. You can add a .rules file (or use the 'Context Servers' feature via MCP) to give Zed persistent project-level instructions. Zed's assistant also reads open files and project context automatically when answering questions."}]},"supportStats":{"totalFeatures":13,"supportedCount":5,"partialCount":2,"notSupportedCount":6,"unknownCount":0,"supportPercentage":46}}],"providers":[{"provider":"Paul Gauthier (open source)","count":1},{"provider":"Anthropic","count":1},{"provider":"Cline Bot Inc. (open source)","count":1},{"provider":"Cursor.sh","count":1},{"provider":"Cognition AI","count":2},{"provider":"Google","count":3},{"provider":"GitHub","count":1},{"provider":"Amazon Web Services","count":1},{"provider":"OpenAI","count":1},{"provider":"anomalyco (open source)","count":1},{"provider":"Replit","count":1},{"provider":"Zed Industries","count":1}],"supportedIDEs":["Antigravity","Antigravity (desktop IDE)","Browser","Cursor","Desktop app","Devin Desktop","Devin Desktop (custom IDE)","Devin Desktop (formerly Windsurf)","Emacs","GitHub (web)","JetBrains","JetBrains (ACP)","JetBrains (Windsurf plugin)","JetBrains (via ACP)","JetBrains IDEs","Jules API","Jules Tools (CLI)","Kiro IDE","Mobile (iOS/Android)","Neovim","Replit (browser)","Replit Desktop","Slack","Terminal (TUI)","Terminal / CLI","VS Code (remote)","VSCode","VSCode (Companion extension)","VSCode (fork)","VSCodium","Visual Studio","Web","Xcode","Zed","Zed (ACP)"],"statistics":{"totalAgents":15,"totalProviders":12,"totalFeatures":13,"averageSupportPercentage":69,"lastUpdated":"2026-10-01T01:57:49.070Z"}},"status":"success"}