How to migrate from Tray.ai to FlowRunner
Export your Tray.ai workflows as JSON, then let FlowRunner Prompt to Flow rebuild each one as a flow. What the export carries, what to finish by hand, and how to switch over.
This guide migrates your Tray.ai workflows to FlowRunner without rebuilding each one by hand. You export each workflow from Tray as a JSON file, and FlowRunner Prompt to Flow reads one at a time and builds the same workflow as a FlowRunner flow, in your workspace.
What’s left for you is the part no export carries: connecting your accounts, adding keys, and testing each flow against what the Tray workflow did. For how the two platforms compare, see FlowRunner vs Tray.ai.
1. Export your workflows from Tray.ai
Tray gives you two exports.
- One workflow. Open the workflow in the builder, click the three dots at the top right of the builder header, and choose Export workflow. Tray saves a JSON file.
- A whole project. Open the Project versions page and export a saved version (Tray recommends this) or the current state. The JSON file holds the project’s workflows, its configuration, solutions, the structure of its data tables and vector tables, API endpoint definitions, and basic agent metadata.
What the exports leave out:
- Authentications. Tray asks you to choose or create them again on import, so they’re never in the file.
- Data table rows and stored vectors. The project file has the table structure, not the data.
- API management policies and roles, and agent links to subprojects.
Tray’s docs put import and export under Enterprise Core and Lifecycle Management and don’t say which plans include them, so check your plan. If you can’t export, describe the workflow in the chat in step 3 and attach a screenshot of each step instead.
Before you switch anything off, copy the input of a few recent runs of each workflow so you have real data to test with in step 5.
2. Connect FlowRunner Prompt to Flow to your AI tool
FlowRunner Prompt to Flow is an MCP server you add to the AI tool you already use. You connect it once and sign in with your FlowRunner account. The server URL is the same for every tool: https://app.flowrunner.ai/mcp.
Claude (web and desktop)
- In Claude, open Settings > Connectors.
- Click + Add, then Add custom connector.
- Enter
FlowRunneras the name andhttps://app.flowrunner.ai/mcpas the server URL. If Claude offers sign-in choices, pick Sign in now and Register automatically, and leave the client ID and secret empty. - Click Add, then Connect, and approve the FlowRunner sign-in page that opens.
- In a chat, click + > Connectors and check that FlowRunner is on.
On Claude's Team or Enterprise plan, an Owner adds FlowRunner once for everyone under Organization settings > Connectors, with the same URL. Each member then finds it under Settings > Connectors and clicks Connect.
Claude Code
-
In your project folder, add the server. Add
--scope userto use it in every project.claude mcp add --transport http flowrunner https://app.flowrunner.ai/mcp -
Sign in, and approve the FlowRunner sign-in page that opens in your browser.
claude mcp login flowrunner -
Check the connection with
claude mcp list. Theflowrunnerline ends in✔ Connected.
ChatGPT
- On the web, open ChatGPT Plugins. ChatGPT connects to MCP servers on chatgpt.com only.
- Click Add, then Add custom MCP server.
- Enter
FlowRunneras the name and, under Connection,https://app.flowrunner.ai/mcpas the server URL. - Choose OAuth, select I understand and want to continue, then Create as a plugin.
- Install the plugin and approve the FlowRunner sign-in page.
- In a chat, type
@and pick FlowRunner.
Whether you can add a custom MCP server depends on your ChatGPT plan and, in a Business, Enterprise or Edu workspace, on your admin's settings.
Codex
-
Add these lines to
~/.codex/config.toml, creating the file if it doesn't exist.[mcp_servers.flowrunner] url = "https://app.flowrunner.ai/mcp" -
Sign in, and approve the FlowRunner sign-in page that opens in your browser.
codex mcp login flowrunner -
Check the connection with
codex mcp list. Theflowrunnerrow showsOAuthin the Auth column.
Cursor
-
Create
~/.cursor/mcp.jsonto use FlowRunner in every project, or.cursor/mcp.jsonin a project folder for that project only. If the file already exists, add theflowrunnerentry inside itsmcpServers.{ "mcpServers": { "flowrunner": { "url": "https://app.flowrunner.ai/mcp" } } } -
Open Customize in the sidebar, then MCPs, and follow the sign-in prompt for
flowrunner.
VS Code (GitHub Copilot)
- Open the Command Palette, run MCP: Add Server, choose HTTP, enter
https://app.flowrunner.ai/mcp, and name itflowrunner. - Open Copilot Chat in agent mode and ask "List my FlowRunner workspaces". The first time, VS Code asks you to sign in: allow it and approve the FlowRunner sign-in page.
- Copilot lists your workspaces. FlowRunner is connected.
On the FlowRunner sign-in page, check the account under Logged in as, tick I recognize and trust this URL, and click Authorize. If you don't have a FlowRunner account yet, the Free plan needs no card, and every new account starts with 14 days of Professional. The setup guide in the docs has every tool step by step.
3. Convert one workflow at a time
Start a new chat, paste this request, then attach or paste one workflow’s JSON after it:
Rebuild this Tray.ai workflow as a FlowRunner flow. It’s the JSON from Tray’s Export workflow: it lists the trigger and each step in order, with its connector, operation, and input settings, plus Boolean Condition and Branch steps, Loop Collection loops, Call workflow steps, and any Script or Python script code. Authentications aren’t in the file. Keep the same behavior. Before you build, tell me anything you can’t map one to one, and ask me about anything the file doesn’t say.
If you exported a project, use this one instead:
This is a Tray.ai project version export. It holds several workflows, the project configuration, and the structure of its data tables, but no table rows and no authentications. Build the callable workflows first as their own flows, list the data tables the project uses so I can decide where their data should live, then build the main workflow. Keep the same behavior, and ask me about anything the file doesn’t say.
Then add a line or two about anything the file can’t tell it:
- who answers each approval step, and whether by email, Slack, WhatsApp, or phone
- whether a workflow saves a date in Data Storage so a second, scheduled workflow can pick it up later (Tray’s pattern for waits past 10 minutes); ask for one flow that waits instead
- which steps use Python script, so it can rewrite them in JavaScript or call your code as a service
- which AI provider and model an agent step used
Answer its questions, and it builds the flow. One workflow per chat keeps the questions specific.
4. Finish what the export can’t carry
Each conversion ends with a short list of what’s left. Expect some of these:
- Connect your accounts. Each app is connected once in the workspace and shared across your flows.
- Add keys and secrets. FlowRunner runs AI agents on your own provider key, added once. Other API keys and secrets go in API Keys too, and an HTTP Request block reads them through a placeholder.
- Set the schedule. Check each scheduled flow’s timing against the Tray workflow’s Scheduled Trigger before you start it.
- Note the new address. A workflow that started from a Webhook Trigger gets a new External Callback URL. That URL includes your workspace key, so treat it like a password, and give it to the calling app when you switch over.
- Check the approvals. The AI sets up each approval request in the channel you named. Two things are added in the flow editor: a Form, if a Tray Form Trigger or custom form collected the answer, and an AI agent that calls another flow as its tool. The AI tells you which of these to add.
- Move table data. Copy the rows of any Tray data table to wherever the AI suggested.
- Start the flow, if the AI left it stopped.
5. Test, then switch over
Run each flow with the inputs you saved in step 1 and compare the results with what Tray produced. Open the run in the flow’s Instances tab to see what each block received and returned. When the results match, point the calling apps and webhooks at FlowRunner, watch the runs for a few days, and disable the Tray workflow.
How Tray.ai maps to FlowRunner
| Tray.ai | FlowRunner | What changes |
|---|---|---|
| Workflow | Flow | One run of a flow is one execution, however many steps it has. |
| Webhook Trigger | External Callback | New URL; it includes your workspace key. |
| Scheduled Trigger | Flow Schedule | Once, interval, daily, weekly, monthly, or cron. |
| Connector operation | Integration action | 2,100+ integrations, all on every plan. |
| Boolean Condition, Branch | Condition, Value Router | The branches show on the canvas. |
| Loop Collection, Break loop | List Iterator, Break | Looping over a list doesn’t add to what a run costs. |
| Callable Trigger, Call workflow, Callable response | Call Flow and Return Result | The called flow can return an answer or run on its own. |
| Data Storage | Shared Memory or Set Variables | Shared Memory keeps key and value data across runs. |
| Delay | Wait | A run waits up to 30 days on Free, Starter, and Growth, and up to a year on Professional and above. |
| Stored date plus a daily scheduled workflow | One flow with a Wait | No second workflow. |
| Script (JavaScript) | Custom Cloud Code | Has fetch; npm packages go in a custom extension. |
| Python script | Custom Cloud Code (rewritten in JavaScript), or your Python service called with an HTTP Request | FlowRunner doesn’t run Python. |
| HTTP Client | HTTP Request | Built-in Retry Policy for status codes and network errors. |
| Approval step, Form Trigger | Human-in-the-loop flow with an External Callback or a Form | On every plan; the person answers by email, Slack, WhatsApp, phone, or a Form. |
| Authentication | App connection | Connected once, shared across flows. |
| Merlin agent | AI Agent | On every plan, on your own provider key. |
Quick wins on day one
- One run, one execution. Tray counts a task for each operation on your data; FlowRunner counts the whole run once.
- Waits that stay in one run. A run waits up to 30 days on Free, Starter, and Growth, and up to a year on Professional and above, where Tray’s Delay step stops at 10 minutes.
- A free plan that stays. Free is $0 for 100 executions a month, with no card, and Starter starts at $5.
- Agents and approvals on every plan. Every integration, AI agents on your own keys, and human-in-the-loop come with every plan, including Free. On Tray, the Merlin Agent Builder is a paid add-on.
Want a hand with a bigger migration? Contact us.
Frequently asked questions
Does FlowRunner import Tray.ai workflow files directly?
No. FlowRunner has no Tray.ai importer. FlowRunner Prompt to Flow, connected to Claude, ChatGPT, Codex, or Cursor, reads the workflow JSON that Tray exports and builds the same workflow as a FlowRunner flow, one workflow per chat. It tells you what it can't map one to one and asks about anything the file doesn't say.
Do my Tray.ai authentications come across?
No. Tray.ai exports don't carry authentications. You connect each account once in FlowRunner, and every flow in the workspace can use it.
What if the Export workflow option isn't in my Tray.ai account?
Tray's docs put import and export under Enterprise Core and Lifecycle Management and don't say which plans include it, so check your plan. If you can't export, describe the workflow to your AI tool step by step and attach a screenshot of each step. FlowRunner Prompt to Flow builds from that the same way.
My Tray workflow waits by saving a date and checking it on a daily schedule. What happens to that?
It becomes one flow that waits. Tray recommends that pattern because its Delay step stops at 10 minutes. In FlowRunner the run itself waits, up to 30 days on Free, Starter, and Growth, and up to a year on Professional and above, so the storage step and the second scheduled workflow go away.
What happens to Python script steps?
FlowRunner doesn't run Python. A short Python function can be rewritten in JavaScript as a Custom Cloud Code step, and FlowRunner Prompt to Flow can do that rewrite. Longer Python code stays where it runs today, as a service the flow calls with an HTTP Request block.
Ready to try it?
Start on the Free plan. No card required, and every new account gets 14 days of Professional.