Fixer
Benchify Fixer is an automated service for fixing common bugs in React apps. The Fixer is much faster and cheaper than an AI coding agent, and has a higher success rate.
Here’s how to use it:
-
Log into Benchify
Log in to your account at app.benchify.com.
-
Navigate to Settings
Click on your profile icon and select “Settings”.
-
Click on Credentials
In the settings menu, select the “Credentials” section.
-
Create a New Key
Click the “Create API Key” button. Give your key a descriptive name.
-
Copy Your New Key
Copy the generated key and save it somewhere secure. You won’t be able to see it again.
For the sake of this document, we will refer to your key as
$BENCHIFY_KEY
. Make sure to replace this placeholder with your actual key when running commands. -
Call the Fixer API and Interpret the Response
Generate a URL pointing to either the public
.git
repository of your (broken) application, or a publicly accessible zip, tar, or tar.gz file containing the source code. For the sake of this document, we will refer to this URL as$REPO_URL
.Hit the Fixer endpoint using your API key and repository URL. You’ll also need to provide a
jobName
(for your reference) and the command needed to build your project (buildCmd
).Here is an example request using a simple (broken) demo we made with Claude Code:
The Fixer will then attempt to fix your code and you’ll receive a response within a few seconds. The response will contain:
build_status
: Indicates whether the build command succeeded (0
) or failed (non-zero) after applying the fixbuild_output
: The console output from running your build commanddiff
: The changes made to fix your code
Here’s another example, from v0, where the Fixer was unable to fully fix the code:
ResponseIf there’s an issue with your request, you’ll receive an error status code with a corresponding message:
Missing Required Parameter (400 Bad Request):
ResponseInvalid API Key (401 Unauthorized):
Response -
Apply the Diff
You can easily apply the diff to your codebase using the
git apply
command. For example:If you are building an AI agent, you may find that even when the Fixer fails, the
diff
it suggests is directionally correct. That is to say, your Agent may want to use thediff
andbuild_output
as a starting point for its own suggestions. (Note, we only bill you for successful Fixer runs, wherebuild_status
is0
, but we give you the full output regardless.)