Turbopack Error: Failed to write app endpoint /page

[mezo2000tiger2-sudo]

Turbopack version: da605e0c
Next.js version: 0.0.0

Error message:

Failed to write app endpoint /page

Caused by:
- Next.js package not found

Debug info:
- Execution of Project::hmr_version_state failed
- Execution of Project::hmr_version failed
- Execution of VersionedContentMap::get failed
- Execution of VersionedContentMap::get_asset failed
- Execution of VersionedContentMap::compute_entry failed
- Execution of *emit_assets failed
- Execution of all_assets_from_entries_operation failed
- Execution of *all_assets_from_entries failed
- Execution of endpoint_output_assets_operation failed
- Execution of <AppEndpoint as Endpoint>::output failed
- Failed to write app endpoint /page
- Execution of AppEndpoint::output failed
- Execution of AppEndpoint::app_page_entry failed
- Execution of *get_app_page_entry failed
- Execution of AppProject::rsc_module_context failed
- Execution of *ModuleAssetContext::new failed
- Execution of *AppProject::get_rsc_transitions failed
- Execution of AppProject::ecmascript_client_reference_transition failed
- Execution of *NextEcmascriptClientReferenceTransition::new failed
- Execution of AppProject::ssr_transition failed
- Execution of *FullContextTransition::new failed
- Execution of AppProject::ssr_module_context failed
- Execution of AppProject::shared_transition failed
- Execution of *FullContextTransition::new failed
- Execution of AppProject::shared_module_context failed
- Execution of *ModuleAssetContext::new failed
- Execution of get_server_resolve_options_context failed
- Execution of get_next_server_import_map failed
- Next.js package not found

[Mehrdoost]

hey @mezo2000tiger2-sudo , ran into this exact stack trace a while back in a monorepo. the Next.js version: 0.0.0 is a dead giveaway that turbopack literally can't find the next package. it's not a code issue, it's node_modules or resolution.

first thing, check if next actually exists:

ls node_modules/next

if it's missing or the folder is empty, that's it. if it's a symlink into some workspace and that workspace isn't built yet, same thing. either build the dep first or just install next directly in the app.

try the nuclear option—fixes this 80% of the time:

rm -rf node_modules .next
rm package-lock.json   # or yarn.lock / pnpm-lock.yaml
npm install

if you're on pnpm, turbopack hates its strict resolution sometimes. in .npmrc add:

shamefully-hoist=true
public-hoist-pattern[]=*next*

then **pnpm install**.

also check if you have multiple versions of next floating around: **npm ls next**. sometimes a mismatched version in a nested dep can mess with the server import map. if the package is there but broken, **npm install --forcel** might rebuild it.

the stack trace bottoms out at get_next_server_import_map. that function literally tries to load next internals and fails cause the package doesn't resolve. so it's not your app code—it's the runtime environment.

if still broken, run the dev server with the local binary directly instead of **npx next dev:**

./node_modules/.bin/next dev

i've seen global npx cache cause weird shadows. let us know if it helps, happy to dig deeper.