Version 7.0 Alpha Upgrade

.gitattributes

@@ -3,5 +3,40 @@
 # For more info see: https://www.aleksandrhovhannisyan.com/blog/crlf-vs-lf-normalizing-line-endings-in-git/
 * text=auto
 docker/** text eol=lf
+
+# The frontend formatters read the file on disk, so a CRLF working tree makes `format:check` fail
+# for every file -- including ones nobody touched. That leaves contributors on Windows unable to run
+# the same check CI runs, which is how formatting errors reach CI in the first place.
+#
+# Only the working copy changes: these files are already stored with LF (`* text=auto` normalizes on
+# commit), so nothing in history is rewritten.
+#
+# It is not automatic on an existing checkout, and nothing warns you. Git applies `eol=lf` when it
+# writes a file, and a pull only writes files whose content changed -- so a CRLF working tree
+# becomes a mixed one: the files this revision touched are LF, the thousands it did not stay CRLF,
+# and `format:check` keeps failing, now on files nobody edited. Rewrite the rest once, from the
+# repository root, with nothing you want to keep uncommitted (this discards working-copy changes;
+# `git stash` first if you have any):
+#
+#     git rm --cached -r .
+#     git reset --hard
+#
+# `rm --cached` empties the index without touching disk; `reset --hard` then rewrites both from
+# HEAD, which is the only step that applies these rules to files whose content did not change.
+# `git status` is clean afterwards on a checkout that was already converted.
+invokeai/frontend/**/*.ts text eol=lf
+invokeai/frontend/**/*.tsx text eol=lf
+invokeai/frontend/**/*.mts text eol=lf
+invokeai/frontend/**/*.js text eol=lf
+invokeai/frontend/**/*.jsx text eol=lf
+invokeai/frontend/**/*.mjs text eol=lf
+invokeai/frontend/**/*.cjs text eol=lf
+invokeai/frontend/**/*.json text eol=lf
+invokeai/frontend/**/*.webmanifest text eol=lf
+invokeai/frontend/**/*.css text eol=lf
+invokeai/frontend/**/*.html text eol=lf
+invokeai/frontend/**/*.md text eol=lf
+invokeai/frontend/**/*.yaml text eol=lf
+invokeai/frontend/**/*.yml text eol=lf
 tests/test_model_probe/stripped_models/** filter=lfs diff=lfs merge=lfs -text
 tests/model_identification/stripped_models/** filter=lfs diff=lfs merge=lfs -text

.github/AGENTS.md

@@ -1,24 +1,11 @@
-# Agent Instructions
-
-## Package Management
-
-This project uses **pnpm** exclusively for package management in the frontend (`invokeai/frontend/web/`).
-
-- ✅ Use `pnpm` commands (e.g., `pnpm install`, `pnpm run`)
-- ❌ Never use `npm` or `yarn` commands
-- ❌ Never suggest creating or using `package-lock.json` or `yarn.lock`
-- ✅ The lock file is `pnpm-lock.yaml`
-
-Use the following pnpm commands for typical operations:
-
-- pnpm -C invokeai/frontend/web install
-- pnpm -C invokeai/frontend/web build
-- pnpm -C invokeai/frontend/web lint:tsc
-- pnpm -C invokeai/frontend/web lint:dpdm
-- pnpm -C invokeai/frontend/web lint:eslint
-- pnpm -C invokeai/frontend/web lint:prettier
-
-## Project Structure
-
-- Backend: Python in `invokeai/`
-- Frontend: TypeScript/React in `invokeai/frontend/web/` (uses pnpm)
+# CI and contributions
+
+- Read affected package guidance before changing checks.
+- Preserve PR, main-push, merge-group, manual-dispatch, and reusable-workflow events. `github.base_ref` is not always populated outside PRs.
+- Change detection must include workflows and relevant shared actions/configuration; test-only and tooling-only changes must trigger owning checks.
+- Preserve tool versions, lockfiles, reproducible installs, hardware/platform matrices, and pinned action revisions.
+- Prefer package completion commands over duplicating them in YAML. For webv2's `pnpm check:release`, install Chromium and required Linux dependencies first.
+- Preserve failure diagnostics with `always()` where appropriate; keep generated review/performance artifacts out of Git.
+- Match hooks to CI's Ruff version/configuration; validation must not rewrite formatting.
+- Follow [the PR template](pull_request_template.md); keep summaries short and omit subagent counts, tiers, and process narration.
+- Workflow changes do not configure hosted branch protection or install local hooks; verify before claiming either.

.github/CODEOWNERS

@@ -27,4 +27,5 @@
 /invokeai/frontend/install @lstein
 /invokeai/frontend/merge @lstein @blessedcoolant
 /invokeai/frontend/training @lstein @blessedcoolant
-/invokeai/frontend/web  @blessedcoolant  @lstein @dunkeroni @Pfannkuchensack @jpphoto
+/invokeai/frontend/webv1  @blessedcoolant  @lstein @dunkeroni @Pfannkuchensack @jpphoto
+/invokeai/frontend/webv2  @blessedcoolant  @lstein @dunkeroni @Pfannkuchensack @jpphoto

.github/actions/install-frontend-deps/action.yml

@@ -1,12 +1,17 @@
 name: install frontend dependencies
 description: Installs frontend dependencies with pnpm, with caching
+inputs:
+  working-directory:
+    description: Directory to install dependencies in
+    required: false
+    default: invokeai/frontend/webv2
 runs:
   using: 'composite'
   steps:
     - name: setup node 22
       uses: actions/setup-node@v6
       with:
-        node-version: '22'
+        node-version-file: '.nvmrc'
 
     - name: setup pnpm
       uses: pnpm/action-setup@v6
@@ -30,4 +35,4 @@ runs:
     - name: install frontend dependencies
       run: pnpm install --prefer-frozen-lockfile
       shell: bash
-      working-directory: invokeai/frontend/web
+      working-directory: ${{ inputs.working-directory }}

.github/pr_labels.yml

@@ -10,7 +10,8 @@ python:
 - changed-files:
   - all-globs-to-any-file: 
     - 'invokeai/**'
-    - '!invokeai/frontend/web/**'
+    - '!invokeai/frontend/webv1/**'
+    - '!invokeai/frontend/webv2/**'
 
 python-tests:
 - changed-files:
@@ -56,4 +57,6 @@ frontend-deps:
 
 frontend:
 - changed-files:
-  - any-glob-to-any-file: 'invokeai/frontend/web/**'
+  - any-glob-to-any-file:
+    - 'invokeai/frontend/webv1/**'
+    - 'invokeai/frontend/webv2/**'

.github/pull_request_template.md

@@ -1,23 +1,29 @@
 ## Summary
 
-<!--A description of the changes in this PR. Include the kind of change (fix, feature, docs, etc), the "why" and the "how". Screenshots or videos are useful for frontend changes.-->
+<!-- Lead with the problem and resulting behavior. Keep this short; show a concrete before/after when useful. -->
 
 ## Related Issues / Discussions
 
 <!--WHEN APPLICABLE: List any related issues or discussions on github or discord. If this PR closes an issue, please use the "Closes #1234" format, so that the issue will be automatically closed when the PR merges.-->
 
 ## QA Instructions
 
-<!--WHEN APPLICABLE: Describe how you have tested the changes in this PR. Provide enough detail that a reviewer can reproduce your tests.-->
+<!-- List actual commands/results and relevant reproduction steps. State skipped, failing, or unavailable checks honestly. Include screenshots/video for changed UI and before/after measurements for material performance claims. -->
 
-## Merge Plan
+## Review
 
-<!--WHEN APPLICABLE: Large PRs, or PRs that touch sensitive things like DB schemas, may need some care when merging. For example, a careful rebase by the change author, timing to not interfere with a pending release, or a message to contributors on discord after merging.-->
+<!-- Summarize material findings resolved and remaining risks or limitations. If there were no material findings, say so briefly. Omit subagent counts, review tiers, and review-process narration. Do not attach review transcripts or work logs. -->
+
+## Compatibility / Rollout
+
+<!-- When applicable: persisted-data migrations, API/generated-artifact compatibility, deployment ordering, and recovery considerations. Omit this section when unnecessary. -->
 
 ## Checklist
 
 - [ ] _The PR has a short but descriptive title, suitable for a changelog_
-- [ ] _Tests added / updated (if applicable)_
-- [ ] _❗Changes to a redux slice have a corresponding migration_
+- [ ] _Meaningful regression coverage added / updated where needed; obsolete tests/code removed_
+- [ ] _Persisted-state and API changes include required migrations / compatibility validation_
+- [ ] _Relevant performance/efficiency opportunities considered; material claims have evidence_
+- [ ] _Material review findings resolved and relevant checks rerun_
 - [ ] _Documentation added / updated (if applicable)_
 - [ ] _Updated `What's New` copy (if doing a release after this PR)_

.github/workflows/build-container.yml

@@ -9,7 +9,7 @@ on:
       - 'invokeai/**'
       - 'docker/Dockerfile'
       - 'docker/docker-entrypoint.sh'
-      - 'workflows/build-container.yml'
+      - '.github/workflows/build-container.yml'
     tags:
       - 'v*.*.*'
   workflow_dispatch:
@@ -100,7 +100,13 @@ jobs:
           password: ${{ secrets.GITHUB_TOKEN }}
 
       - name: Build container
-        timeout-minutes: 40
+        # The rocm image's venv layer is ~7.2GB compressed - roughly 60% larger than
+        # cuda's and 9x cpu's - and BuildKit gzips a single layer on one thread, so
+        # the export alone is ~6.5 min on a healthy runner. The whole job normally
+        # lands at ~17 min, but every phase roughly doubles on a slow runner, which
+        # blew through the old 40 min cap four times (2026-07-28, 2026-09-17 x2,
+        # 2026-09-22).
+        timeout-minutes: 90
         id: docker_build
         uses: docker/build-push-action@v7
         with:
@@ -109,7 +115,21 @@ jobs:
           platforms: ${{ env.PLATFORMS }}
           build-args: |
             GPU_DRIVER=${{ matrix.gpu-driver }}
-          push: ${{ github.ref == 'refs/heads/main' || github.ref_type == 'tag' || github.event.inputs.push-to-registry }}
+          # `outputs` rather than `push` so we can set the compression level: BuildKit
+          # defaults to gzip -6, which runs at ~19MB/s over that one big layer. gzip -1
+          # is 2-3x faster for ~15% larger layers, which is a good trade when the
+          # bottleneck is CI wall-clock rather than registry bandwidth. An empty
+          # `outputs` emits no --output at all, which is what a non-pushing build did
+          # before.
+          #
+          # The `== 'true'` is load-bearing. `github.event.inputs.*` is always a string,
+          # and every non-empty string is truthy, so a bare reference to an unchecked
+          # boolean input evaluates as TRUE. `push:` got away with it because the action
+          # ran that input through core.getBooleanInput, which maps "false" to false;
+          # `outputs:` is read with getInputList and passes any non-empty string straight
+          # through to --output. Without the comparison, a manual run from a non-main
+          # branch with the box unchecked would publish that branch to ghcr.
+          outputs: ${{ (github.ref == 'refs/heads/main' || github.ref_type == 'tag' || github.event.inputs.push-to-registry == 'true') && 'type=image,push=true,compression=gzip,compression-level=1' || '' }}
           provenance: mode=max
           sbom: true
           tags: ${{ steps.meta.outputs.tags }}

.github/workflows/deploy-docs.yml

@@ -78,12 +78,12 @@ jobs:
         with:
           version: '0.11.28'  # When updating this, make sure the version in launcher's package.json gets updated too.
           enable-cache: true
-          python-version: '3.11'
+          python-version: '3.12'
 
       - name: setup python
         uses: actions/setup-python@v6
         with:
-          python-version: '3.11'
+          python-version: '3.12'
 
       # generate_docs_json.py only needs the invokeai package importable
       # (pydantic + invokeai.app/backend). Skip the [test] extra to keep CI fast.
@@ -107,6 +107,10 @@ jobs:
         working-directory: docs
 
       # Checks
+      - name: test docs tooling
+        run: pnpm test
+        working-directory: docs
+
       - name: verify generated docs data
         run: pnpm run check-docs-data
         working-directory: docs

.github/workflows/frontend-checks.yml

@@ -30,67 +30,49 @@ on:
         type: boolean
         default: true
 
-defaults:
-  run:
-    working-directory: invokeai/frontend/web
+# One run per pull request: a new push cancels the superseded run. The PR number is used rather
+# than the head branch, which two forks can share. Pushes to main and merge groups fall back to
+# the run id and never cancel each other. The group names the workflow literally
+# because `github.workflow` is the caller's name inside a reusable workflow, and release.yml calls
+# several of these at once.
+concurrency:
+  group: frontend-checks-${{ github.event.pull_request.number || github.run_id }}
+  cancel-in-progress: true
 
 jobs:
-  frontend-checks:
+  frontend-webv2-checks:
     runs-on: ubuntu-latest
     timeout-minutes: 10 # expected run time: <2 min
+    defaults:
+      run:
+        working-directory: invokeai/frontend/webv2
     steps:
       - uses: actions/checkout@v6
 
-      - name: Fail if package-lock.json is added/modified (pnpm only)
-        shell: bash
-        working-directory: .
-        run: |
-          set -euo pipefail
-          git fetch --no-tags --prune --depth=1 origin "${{ github.base_ref }}"
-          if git diff --name-only "origin/${{ github.base_ref }}...HEAD" | grep -E '(^|/)package-lock\.json$'; then
-            echo "::error::package-lock.json was added or modified. This repo uses pnpm only."
-           exit 1
-          fi
-
-      - name: check for changed frontend files
+      - name: check for changed webv2 files
         if: ${{ inputs.always_run != true }}
         id: changed-files
-        # Pinned to the _hash_ for v45.0.9 to prevent supply-chain attacks.
-        # See:
-        # - CVE-2025-30066
-        # - https://www.stepsecurity.io/blog/harden-runner-detection-tj-actions-changed-files-action-is-compromised
-        # - https://github.com/tj-actions/changed-files/issues/2463
         uses: tj-actions/changed-files@9426d40962ed5378910ee2e21d5f8c6fcbf2dd96
         with:
           files_yaml: |
-            frontend:
-              - 'invokeai/frontend/web/**'
+            webv2:
+              - '.github/workflows/frontend-checks.yml'
+              - '.nvmrc'
+              - '.github/actions/install-frontend-deps/**'
+              - 'invokeai/frontend/webv2/**'
 
       - name: install dependencies
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
+        if: ${{ steps.changed-files.outputs.webv2_any_changed == 'true' || inputs.always_run == true }}
         uses: ./.github/actions/install-frontend-deps
+        with:
+          working-directory: invokeai/frontend/webv2
 
-      - name: tsc
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
-        run: 'pnpm lint:tsc'
-        shell: bash
-
-      - name: dpdm
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
-        run: 'pnpm lint:dpdm'
-        shell: bash
-
-      - name: eslint
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
-        run: 'pnpm lint:eslint'
-        shell: bash
-
-      - name: prettier
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
-        run: 'pnpm lint:prettier'
+      - name: lint
+        if: ${{ steps.changed-files.outputs.webv2_any_changed == 'true' || inputs.always_run == true }}
+        run: 'pnpm run lint'
         shell: bash
 
-      - name: knip
-        if: ${{ steps.changed-files.outputs.frontend_any_changed == 'true' || inputs.always_run == true }}
-        run: 'pnpm lint:knip'
+      - name: build
+        if: ${{ steps.changed-files.outputs.webv2_any_changed == 'true' || inputs.always_run == true }}
+        run: 'pnpm run build'
         shell: bash