Give Replit Agent access to FetchSandbox, prove the link with a real tool call, then verify your running app against isolated provider twins. The twins need no live vendor keys.
A FetchSandbox account key from https://fetchsandbox.com/keys. Access to Replit Integrations, which may depend on your workspace permissions or feature availability.
01
Get your account key
Copy your key from https://fetchsandbox.com/keys. Keep it private and never paste real keys into chat.
Open Settings > Integrations > Your integrations > Add custom > Add personal MCP server. Name it FetchSandbox and set the URL to https://fetchsandbox.com/mcp/v1. In Advanced settings, add a custom header named Authorization with the value Bearer followed by your own FetchSandbox key. Then click Test & save. Source: https://docs.replit.com/build/connect-via-mcp.
Replit documentation screenshot · reviewed 12 Nov 2026. Setup reference. This shows the platform setup, not an execution.
03
Confirm an actual tool call
Mention FetchSandbox in Agent chat and ask Agent to call list_workflows with spec_slug=stripe. This is a read-only connection test. A list of available tools doesn't count, because tool discovery is not workflow proof.
Confirm an actual tool call
Call FetchSandbox list_workflows with spec_slug stripe and tell me the actual result.
04
Verify the running app and inspect a receipt
Ask Agent to run validate_integration. Agree on the observable behavior, configure your app's calls to use the supplied twins, invoke the app and inspect the receipt. Local apps use managed tunnel transport. Hosted apps need a reachable protected URL. Connecting the MCP to Agent does not automatically send your app's provider calls to the twins.
Verify the running app and inspect a receipt
Use FetchSandbox validate_integration for this running app. Agree the observable expected behavior with me, configure the isolated twins, invoke the actual app and report the receipt, including anything failed or unmeasured. Keep polling the same get_job until it reaches a terminal state.
Signs your setup works
Agent shows a real list_workflows result, and your running app produces a validate_integration receipt.
Scope of this evidence
The receipt records expected versus actual for the agreed scenarios, tested against your configured twins, not live vendors. Point your app's provider calls at the twins yourself, because the MCP connection doesn't reroute them. A reference quickrun doesn't test your code. Related public receipts illustrate workflows. They're separate from verifying your app here and aren't native Replit acceptance. Keep your key in the Authorization header, never in chat.
Historical runs recorded against isolated twins. You can inspect each receipt and its evidence. Not a new execution or production certification.
Paid order provisioning (Stripe events)
Passed
Sales needs to grant product access the moment an order is paid, and never for an order that was not paid in full, was refunded or disputed, or whose payment event was forged, replayed, delayed or lost.
Each receipt is one individual dated run against isolated provider twins, not live vendors. The caller declared the app revision, and it is not attested. Workflow docs are pinned to the commit shown here. A status appears only when the original verdict, counts, completeness and app_verified all support it, and counts are never merged. Some source exports, including the n8n lead flow, are not public. These public examples are separate from proving your own app through your builder, and they don't show that each native client was tested.
Customer Experience needs to acknowledge cancellation requests, evaluate subscription terms, obtain exceptions where required, and let the correct customer choose an eligible refund method.
Each receipt is one individual dated run against isolated provider twins, not live vendors. The caller declared the app revision, and it is not attested. Workflow docs are pinned to the commit shown here. A status appears only when the original verdict, counts, completeness and app_verified all support it, and counts are never merged. Some source exports, including the n8n lead flow, are not public. These public examples are separate from proving your own app through your builder, and they don't show that each native client was tested.
