# AI coding assistants: connect your agent to Deplomo Connect supported AI coding assistants, including Claude Code, Cursor, and Codex, to Deplomo with one command. Explore the available service-management tools; service creation and deployment are currently paused. [Choose your agent](#directory)[MCP connection guide](/support/docs/cli/agents) Give this prompt to your agent (Claude Code, Cursor, Codex or another supported client): `Set me up on Deplomo: https://deplomo.com/agent-ready.md` Claude Inspect this project and tell me what it would need before deployment. Company website project ## Connect in three steps Set it up once to use it across projects on this device. ### Install and sign in Install the Deplomo CLI and sign in with deplomo login. ### Connect an agent The install command adds the MCP connector and Deplomo guide to your project. ### Describe the outcome you want Tell the agent what you want to do. It chooses the tools and checks the result. ## Choose the tool you code with Each tool page shows the setup command and config file path, plus example prompts to get started. [### Codex OpenAI agent Connect Codex to Deplomo MCP tools to inspect existing service status and logs, where available. Service creation and deployment are currently paused. `.codex/config.toml`](/agents/codex) [### Claude Code Anthropic agent Connect Claude Code to Deplomo MCP tools to inspect available information about existing services. Service creation and deployment are currently paused. `.mcp.json`](/agents/claude-code) [### Cursor Smart code editor Connects Cursor's built-in agent to live project and service data from Deplomo, so its answers reflect the actual state of your infrastructure. `.cursor/mcp.json`](/agents/cursor) [### Gemini Antigravity Google agent Connects to Deplomo resources through MCP to perform cloud tasks from within Antigravity. `.agents/mcp_config.json`](/agents/antigravity) [### GitHub Copilot Agent mode in VS Code Connects Copilot agent mode to Deplomo through VS Code's MCP configuration file. `.vscode/mcp.json`](/agents/copilot) [### OpenCode Open-source agent Enables Deplomo as a local MCP server in OpenCode. The choice of model is still up to you. `opencode.json`](/agents/opencode) [### Devin Software development agent Connect Devin to Deplomo MCP tools to inspect available service information and diagnostics. Service creation and deployment are currently paused. `.devin/mcp_config.json`](/agents/devin) [### Pi Command-line agent Provides a local connection for working with Deplomo resources from within a command line workflow. `.pi/mcp.json`](/agents/pi) [### Grok xAI agent Uses the shared MCP template to view project status and perform authorized operations on Deplomo services. `.mcp.json`](/agents/grok) ## Compare AI coding assistants and software development agents The table compares example tools by environment, workflow, and MCP support. Model names are illustrative; availability and context limits depend on the selected model and provider.

AI assistant	Environment	Example models	Typical features	Setup guide
Claude Code	Terminal CLI	Claude 3.7 / 3.5 Sonnet	Terminal-based coding assistance for tasks such as code analysis and debugging	Claude Code
Cursor	Code editor based on VS Code	Examples include Claude Sonnet and GPT-4o; availability varies by plan and provider	AI-assisted editing across files, code chat, and frontend coding support	Cursor
OpenAI Codex / CLI	Terminal and shell assistant	Examples include the o-series and GPT-4o; availability varies by plan and provider	Help with system scripts and Linux command troubleshooting	Codex
Gemini Antigravity	Google’s agent-focused development environment	Gemini 2.0 Pro / Flash	Multi-agent workflows and architecture review; context limits depend on the selected model and provider	Gemini Antigravity
GitHub Copilot	VS Code and JetBrains extensions	Examples include GPT-4o and Claude; availability varies by plan and provider	Code completion and coding assistance in the editor	GitHub Copilot
Devin AI	AI software engineering agent	Cognition models and tools	Can assist with issue triage, tests, pull requests, and end-to-end testing, depending on setup	devin ai
OpenCode	Open-source terminal environment	Local models such as Ollama or cloud APIs	Model choice and customization; data handling depends on the selected model and provider	OpenCode
Pi Coding Agent	Lightweight terminal coding agent	Optimized coding models, depending on configuration	Can process commands, inspect logs, and assist with routine tasks, subject to permissions	pi coding agent
Grok (xAI)	Analysis and problem-solving agent	Grok 2 / Grok 3	Can help analyze complex log patterns and longer software context; results depend on model and input	Grok setup

## A guide to choosing an AI coding assistant and connecting it to cloud tools AI coding tools have changed how many software teams work. Coding assistants can help explore a repository, run tests, and work with infrastructure tools when configured and authorized. What an agent can do depends on its model, connected tools, and permissions. Service creation and deployment through Deplomo are currently paused. ### The Model Context Protocol (MCP) for connecting AI agents to cloud tools MCP provides a standard way to connect compatible agents to external tools. After configuring an agent, you can use the Deplomo tools available to that client and account, subject to permissions. These tools can provide information such as service status, logs, metrics, or domains; do not share secret values. Service creation and deployment are currently paused. ### MCP tools for PaaS services With an MCP client configured, you can ask about available information for a Deplomo service without switching to the web dashboard. The tools can provide details such as service status, build or runtime logs, and metrics, depending on access. [PaaS service](/products/paas) Service creation and deployment are currently paused. Review any changes suggested by your coding assistant before applying them. ### Choosing an AI coding assistant for your project The right tool depends on your workflow and the features available in your chosen model and plan: - If you spend most of your time in an editor, tools such as Cursor and GitHub Copilot offer editor-based coding assistance. Features and responsiveness depend on your setup. - For terminal-based work, Claude Code and OpenAI Codex provide command-line workflows; their capabilities depend on the model, configuration, and permissions. - For multi-step tasks and codebase review, tools such as Devin and Gemini Antigravity offer agent-based workflows. Results depend on the task and configuration. ### Data handling when connecting agents to cloud tools When an agent is connected through MCP, it can use the tools available to that client and account, subject to permissions. Review each requested action and avoid sharing secret values in prompts. Available operations vary; service creation and deployment through Deplomo are currently paused. ## The agent sees the current service state MCP connects your agent to Deplomo. The agent uses it to read account information and work in the specified project and environment. The installed Deplomo guide explains the checks and tool limits. ### Project and service The agent finds the right project and applies changes to the intended service. ### Deploy code The agent deploys the current folder and waits for the build result. ### Logs and usage metrics The agent compares build and runtime logs with real metrics to diagnose errors. ### Domains and access The agent connects your domain to the service and shows the DNS records you need. ### Environment variables The agent identifies required variables. It sets secret values only with your approval. ### Usage and costs The agent reads usage and billing information directly from your Deplomo account. ## Try it on a real project “Deploy this project to Deplomo. If the build fails, read the logs, diagnose the cause and tell me before changing any settings.” [Create a Deplomo account](https://panel.deplomo.com/signup)[Read the connection guide](/support/docs/cli/agents)