[hidb4ai]

Which issue or RFC does this PR close?

Release preparation for the merged PRs labeled 1.5.1. No issue is closed by this version update.

Rationale for this change

Prepare PowerContext 1.5.1 with consistent installation references, generated contracts and bilingual release notes. Changing the OpenAPI version also changes Desktop's contract digest, so the new digest needs its own tested compatibility record.

What changes are included in this PR?

• Run make version-bump VERSION=1.5.1 and regenerate the Python, Desktop and Dify contract artifacts.

• Include migration target documentation in version automation, with regression coverage that preserves frozen schema revision identities. Update the documented chain through v1_5_0_004 without changing migration resources or upgrade policy.

• Add English/Chinese website notes and publication-ready notes in release/notes/powercontext-v1.5.1.md, based on merged release work, including capture coverage feat(capture): declare bounded projection coverage #2451. Exclude closed/unmerged work and changes already present in 1.5.0; retain experimental and upgrade-support boundaries.

• Add a Desktop compatibility record for Server source bcc0d93a202396340d3724a13def1efffb2223b3 and the exact tested wheel. Preserve all 62 historical records and document the protocol evidence and limits in desktop/VALIDATION.md.

• Refresh all six English/Chinese README GIFs from the current release code in the application light theme, using synthetic Atlas Commerce records. Update the English, Chinese and Japanese captions. Dashboard recordings use Chromium; Desktop recordings use the actual Tauri/WebKitGTK native application connected to the same isolated Server.

Are there any user-facing changes?

Installation instructions, API release metadata and the website changelog now target 1.5.1. This PR adds no runtime API behavior or database schema change. Independent Desktop/plugin package versions and frozen migration numbers remain unchanged. README product walkthroughs now show the current light interface.

The policy remains upgrade-required; these preparation checks do not certify existing-database upgrades. The stable distribution version will come from powercontext-v1.5.1. No tag, GitHub Release, package publication or production deployment is created by this PR.

How was this change tested?

• make release-check VERSION=1.5.1

• make contract-test: 67 passed.

• uv run pytest tests/test_release_version.py tests/test_ci_release_verification.py tests/test_upgrade_release_gate.py -q: 164 passed, 1 skipped (the non-Linux process-ownership case).

• make check and uv run prek run -a: passed, including 61 integration-manifest tests, lock consistency and type checks.

• pnpm --dir desktop generate:check and uv run python integrations/dify/generate_contract.py --check

• make docs-test: passed, including website tests, source-link validation, TypeScript, the production build, and static-export verification of 1,109 public pages and their internal links.

• make build and Twine metadata checks for both wheel and sdist. Installed the VCS-versioned wheel in an isolated environment, ran dependency checks, and passed scripts/ci_release_smoke.py for CLI, Server, MCP and SQLite Memory operations. All 745 runtime files match the exact source commit.

• Desktop's unchanged native product probe passed anonymous, Bearer, HTTPS CA/base-path and enforced authentication Provider modes against real HTTP/SQLite Servers. A standalone wrapper uses the actual product modules and 307 dependencies matching the product lock; the host itself lacks full Tauri/WebKitGTK build prerequisites (the separate README capture below uses a Debian container). After default startup waits timed out, the complete protocol run used a 60-second Server readiness wait, with business assertions and request timeouts unchanged. This does not establish the default startup budget, installed GUI/platform acceptance or signed distribution qualification.

• README capture validation: built the current native Tauri application in a Debian container and exercised Scope selection, full-text search and exact revision reading through native IPC and a real HTTP/SQLite Server. Independently read both displayed memory revisions back from the public API. Fully decoded all six GIFs, checked looping, dimensions, frame brightness and image references, and inspected representative frames. Re-ran uv run prek run -a. These Linux development captures do not establish Windows/macOS packaged acceptance.

The full Python/platform and installed-release database upgrade matrices remain CI/release gates and were not executed locally.

AI usage statement

OpenAI Codex assisted with version synchronization, release-note research and writing, generated artifacts, regression coverage, protocol qualification, light-theme UI recordings and validation.