Docs

Pull requests

Sypnia Cloud only: how Sypnia opens a pull request with the fix, what it contains, what the GitHub App is allowed to do, what it never does, and the limits per plan.

Applies to Cloud At launch

View as Markdown
On this page
  1. Before you start
  2. When a pull request can be opened
  3. What happens
  4. Sypnia reads the repository
  5. It writes the change and a test
  6. It opens the pull request
  7. Your checks run, a person reviews
  8. After the merge
  9. What the pull request contains
  10. What the GitHub App may do
  11. What it never does
  12. When it cannot
  13. Limits per plan

Cloud On Sypnia Cloud, a problem can arrive with its fix as a pull request. Sypnia writes the change and a test on a branch of your repository and opens the pull request. A person reads it and merges it. Sypnia never merges.

The open-source engine does not open pull requests. There, hand the fix to your own coding agent with the prompt or the MCP server. It is the same information.

Before you start

  • The project is connected to a repository through the Sypnia GitHub App. See The GitHub App for how to install it and choose repositories.
  • Your app sends its release as the git commit, so Sypnia knows which code was running. See Releases and source maps.

When a pull request can be opened

ConditionWhy
The analysis is ready with confidence high or mediumSypnia does not write a fix it cannot justify. With low, or not_enough_evidence, the button is off and says why
The problem is openA resolved or ignored problem needs no fix
The problem has no open pull request from SypniaOne at a time per problem
The organization has fixes left this monthSee the limits below

Then it opens in one of two ways:

  • You ask. The button «Open pull request» on the problem, or POST /v1/problems/{problem}/pull-request.
  • By itself, if you turn it on: the project setting agent.auto_pull_requests set to high_confidence opens one for every analysis with confidence high. It is off by default.

What happens

  1. Sypnia reads the repository

    Through the GitHub App, at the commit of the release where the problem happens. It reads the files the analysis names and the ones around them. Your code is used for this pull request and is not kept afterwards.

  2. It writes the change and a test

    The smallest change that removes the cause, and a test that fails without it, in your project’s own test framework. If the project has no tests, the pull request has the change only and says so.

  3. It opens the pull request

    On a new branch named sypnia/<problem slug>, for example sypnia/pay-button-does-nothing, against your default branch.

    You should see
    Within a few minutes the problem shows the pull request, and the problem.pull_request_opened webhook is sent.
  4. Your checks run, a person reviews

    Sypnia does not run your code. Your own CI runs on the pull request, as for any other. You read it, ask for changes or change it yourself, and merge.

  5. After the merge

    When the fix is deployed, mark the problem resolved, or let your coding agent do it. If the failure comes back, the problem reopens by itself.

What the pull request contains

PartContent
TitleWhat the change does, in the imperative: «Read the cart total from its new field»
DescriptionThe problem’s sentence, who it hit and since when; the cause with its evidence; what was changed and why; how the test covers it; links to the problem and to a session with a replay; the confidence
CommitsOne for the change, one for the test, authored by sypnia[bot]
FilesOnly the files the fix needs. Never workflow files, lock files, or configuration of your CI

The description never contains names, e-mails or ids of the people affected.

What the GitHub App may do

The app asks for three permissions, on the repositories you choose and no others.

PermissionAccessWhat it is for
ContentsRead and writeRead the code; push the branch sypnia/<problem slug>
Pull requestsRead and writeOpen the pull request and update its description
MetadataReadRequired by GitHub for any app: the repository’s name and default branch

What it never does

  • It never merges a pull request, and never approves one.
  • It never pushes to your default branch or to any branch it did not create. It never force-pushes over someone else’s commits.
  • It never changes workflows, secrets, settings, collaborators or branch protection. It has no permission to.
  • It never opens more than one pull request for a problem at a time, and never reopens one you closed.
  • It never runs your code or your tests.

Protect your default branch on GitHub so that every pull request needs a review. Sypnia’s pull requests then follow the same rule as anyone’s.

When it cannot

What you seeWhy
The button is off: «Not enough evidence to write a fix»The analysis has no cause, or its confidence is low
The button is off: «Connect a repository»The project has no repository through the GitHub App
failed, with a reason, on the problemSypnia read the code and could not write a change it trusts: the code at that commit does not match the analysis, or the fix would touch more than it should. Nothing was opened and it did not count against your limit
The pull request is open and your checks failSypnia does not retry by itself. Fix it on the branch, or close it and ask again

Limits per plan

A pull request counts when it is opened. Asking again for the same problem counts again. The count is per organization, per calendar month in UTC.

PlanPull requests Sypnia opens a month
Free5
Builder50
Team300

These numbers come from the plans and are shown, with what you have used, by GET /v1/usage. At the limit the button says when the count resets; the analysis and the prompt keep working.

Sections

Start hereWhat Sypnia is, how it works, what is open source and what is the paid cloud, and the two ways to get your first problem in plain words.ConceptsThe ten ideas Sypnia is made of, each defined precisely, from a project and its keys to problems, traces and alerts.SDKsHow your app sends data to Sypnia. One SDK for the browser, one for Node, one for Next.js, and the OpenTelemetry intake for every other backend.The agentHow Sypnia analyses a problem, what it writes, what it sends to a model and what it never does.MCP and skillsConnect your coding agent to Sypnia so it can ask what is failing in production, read what it needs to fix it, and set Sypnia up for you.API referenceThe Sypnia HTTP API: one API for the open-source engine and for Sypnia Cloud, with every endpoint, parameter, response and error.CLI referenceThe sypnia command: how to install it, the variables it reads, what its exit codes mean, and where each command is documented.Self-hostingWhat you run when you host the Sypnia engine yourself, what you need, what it costs, and what you do not get compared with Sypnia Cloud.How the engine is builtThe whole Sypnia engine on one page, from a click in your app to the sentence your coding agent reads, with the design rules and the code that holds them.Sypnia CloudHow the paid service works, from organizations and roles to plans, limits, billing, where your data lives, the GitHub App and support.Privacy and securityWhat Sypnia collects from your app and its users, what it never collects, how masking works, how to delete data, and how the service is protected.LicensingWhich licence covers each part of Sypnia, what you may do with it, how the name may be used, and the terms of a contribution.GuidesStep-by-step recipes for the common ways to use Sypnia, from apps made with AI builders and Next.js on Vercel to Supabase, source maps in CI, alert e-mails from your own engine, and moving from Sentry.ReferenceThe pages you look things up in, from the event schema and the glossary to what is planned, the changelog and the common questions.
↑↓ to move↵ to openEsc to close