Integration
Deploy from the assistant you already build with.
Claude, ChatGPT, Cursor and Codex connect to GetFlowing through one MCP connector. The assistant reads what an app needs to deploy, checks the code, brings the app in and deploys it, and when a deploy fails it reads why and fixes the code. GetFlowing builds, scans, deploys, monitors and keeps the app running.
Integration
AI assistants
How it works
Add the GetFlowing connector to your assistant with the address shown in GetFlowing under Settings, Assistants, sign in to GetFlowing and choose what the assistant may do. Then ask it to deploy. It works as you and only within what you allowed: it can read your apps, follow a build, read logs, deploy and ask for a release, and anything your workspace holds for approval still waits for a person.
The path
From a chat to a running app.
The same path whichever assistant you use. Each step is a tool the connector gives the assistant, so it does not have to guess how GetFlowing works.
Connect
Add the connector to Claude, ChatGPT, Cursor or Codex, sign in to GetFlowing, and tick what the assistant may do: read, chat, bring code in, deploy or approve.
Build to the requirements
The assistant reads the deploy requirements for your plan, or starts from a starter app that already deploys: a web server on the port GetFlowing gives it, a page it answers when GetFlowing checks its health, and a Dockerfile, which GetFlowing writes itself for a stack it recognises.
Check
Before anything is imported, built or charged, the connector checks the code and says what must change, including the dependency audit to run and the secrets to keep out of it.
Bring it in
Import the app from GitHub, or, where your workspace offers it, let the assistant send the code itself. Either way it becomes a managed app, and later versions arrive the same way or with a push.
Deploy
GetFlowing builds the container, scans it and deploys it. Where your workspace asks before each deploy of a released app, the release and every deploy after it wait for a person to approve.
Fix
If a deploy fails, the assistant reads the failure and the logs through the connector, fixes the code they point at and sends a new version.
Specification
The connector, stated plainly.
Every assistant uses the same connector address and signs in to GetFlowing the same way. GetFlowing shows the address and these steps under Settings, Assistants, and deploy instructions you can keep in your project.
- Assistants
- Claude (claude.ai, the desktop app and Claude Code), ChatGPT, Cursor and Codex. Another assistant that can add a remote MCP server with OAuth sign-in can use the same connector.
- Claude Code
- In the project folder, run claude mcp add --transport http getflowing followed by the connector address, then run /mcp and sign in. Then ask it to deploy: the connector tells it what an app needs. Keeping the deploy instructions in CLAUDE.md is optional.
- Claude
- On claude.ai or in the desktop app, open Customize, Connectors, Add, then Add custom connector, and paste the connector address. On Team and Enterprise plans an owner adds it once under Organization settings, Connectors, then each person connects.
- ChatGPT
- On chatgpt.com, open Plugins, choose Add custom MCP server, paste the connector address as the Server URL with OAuth, create it as a plugin, then install it and sign in. In Business, Enterprise and Edu workspaces an admin or owner may need to set it up. Which ChatGPT plans can add a custom MCP server is up to OpenAI.
- Cursor
- Use the Add to Cursor link in GetFlowing, or put the entry in ~/.cursor/mcp.json. Cursor asks you to sign in the first time it uses the connector. Keeping the deploy instructions in AGENTS.md is optional.
- Codex
- Run codex mcp add getflowing --url followed by the connector address, then codex mcp login getflowing. Keeping the deploy instructions in AGENTS.md is optional.
- Sign-in and permissions
- You sign in to GetFlowing yourself, see which workspace and account you are connecting, and choose what the assistant may do. Each connection is listed under Settings, Assistants, where you or a workspace owner can revoke it.
- What it can do
- Read your apps, builds, logs, security findings and usage; check code and bring it in where you allowed it; deploy and ask for a release; and, on plans with GetFlowing’s agent, hand work to the agent in an app’s chat.
- What it cannot do
- Billing, secrets, members, deleting an app and taking it offline stay in GetFlowing. An assistant never approves anything by itself: where the workspace owner allows approving through an assistant, each approval needs your own confirmation.
- Security
- The connector tells the assistant to make the app safe before sending it: audit the dependencies and upgrade any with a high or critical vulnerability, commit the lockfile, keep secrets out of the code and run as a user other than root. GetFlowing scans every app too, and the workspace owner decides whether findings block a deploy.
- Plans
- Every plan can connect an assistant. On Deploy, the assistant makes the changes and GetFlowing deploys them, and your first app can run free for 3 days before you pay. On plans with GetFlowing’s agent, the assistant can also hand work to the agent.
Private beta
Bring software delivery into one system.
Build something new or take control of what already exists. Get the planning, engineering, deployment, and operational capability of an experienced software team.