OpenAPI import ignores the target branch (--branch / targetBranchId) and writes to main

[Pruthvik007]

Summary

When importing an OpenAPI file into a sprint branch, the import ignores the specified branch and writes all resources to the main branch instead. The target sprint branch remains empty.

This happens with both the CLI (apidog import --branch <name>) and the REST API (POST /v1/projects/{id}/import-openapi with targetBranchId). It occurs whether the branch is specified by name or by numeric id, and it persists with External AI Edit Permissions enabled.

Importing through the desktop UI (with the branch selected in the top-left) works correctly. Only the programmatic paths are affected, which means there is currently no way to import a spec into a sprint branch from CI.

This looks like a regression or gap in issue #112 ("Sprint Branch supports import/export").

Environment

• apidog-cli version 2.2.8 (current latest).
• Also reproduced directly against the REST API (api.apidog.com, header X-Apidog-Api-Version: 2024-03-28), so the behaviour is server-side rather than CLI-specific.

Steps to reproduce

1. Create an empty sprint branch:

   apidog branch create --project <projectId> --type sprint --name feat-x --from main
   

2. As a control, create a single endpoint on that branch. This works correctly and isolates to the branch:

   apidog endpoint create --project <projectId> --branch feat-x --file one-endpoint.json
   

   Result: the main branch endpoint count is unchanged; the sprint branch gains one endpoint.

3. Import an OpenAPI file into the same branch:

   apidog import --project <projectId> --branch feat-x --format openapi --file spec.yaml
   

   Result: the main branch endpoint count increases; the sprint branch remains empty.

4. Confirm where the resources landed:

   apidog endpoint list --project <projectId> --branch main     # contains the imported endpoints
   apidog endpoint list --project <projectId> --branch feat-x    # returns 0
   

Observed results

Using a spec containing 3 endpoints and 2 schemas, imported into a newly created, empty sprint branch:

Action	Resources added to main	Resources added to the sprint branch
endpoint create --branch feat-x (control)	none	the single endpoint (isolated correctly)
import --branch feat-x	all 3 endpoints and 2 schemas	none
REST import-openapi with targetBranchId	all imported resources	none
import --branch <numeric branch id>	all imported resources	none

The important detail is that endpoint create --branch isolates correctly while import --branch does not, using the same flag, the same branch, and the same default module. This rules out branch resolution, module, and permissions as the cause, and points to the import operation specifically.

Expected behaviour

import --branch <name> and the REST import-openapi call with targetBranchId should place the imported endpoints and schemas on the specified sprint branch, leaving the main branch unchanged, as endpoint create --branch and the UI import already do.

Actual behaviour

All imported resources are written to the main branch, and the target sprint branch stays empty.

Impact

This blocks any CI or pipeline workflow that imports a spec into a per-feature sprint branch, for example to keep a feature branch's mock and documentation isolated from main.