ContextVault v0.2.1 — Export Recovery and Build Stability Hotfix

ContextVault v0.2.1 is a cumulative reliability hotfix for the Windows portable release. The archive format and schema remain unchanged at 1.0.

Export recovery

• Recovers virtualized conversation scans when message semantics are stable but the surrounding ChatGPT DOM continues mutating.
• Distinguishes bounded stalled-image recovery from general DOM/loader recovery.
• Uses visible blocking-loader evidence, semantic stability, forced upward scrolling, and bounded non-empty reload recovery without discarding verified checkpoints.
• Improves scroll-container selection for nested ChatGPT layouts and reports exact blocker/scroller/recovery diagnostics on failure.
• Retries transient Playwright execution-context/DOM observation failures in place.

Capture integrity

• Preserves mixed conversation-turn-*, data-message-id, and role-based message container families.
• Reads nested source message identity and author role from wrapper structures.
• Includes export-critical image, attachment, and file resource attributes in semantic stability signatures.
• Treats checkpoint storage or round-trip failures as infrastructure failures rather than silently degrading valid messages.

Workflow lifecycle

• Releases the exclusive managed-browser lease inside the worker lifecycle before failed/completed/cancelled task state becomes externally terminal.
• Keeps an idempotent Future callback fallback for tasks cancelled before execution starts.
• Prevents delayed callbacks for an old lease from releasing a newer workflow owner.

Portable runtime and packaging

• Validates packaged application-root candidates against the complete runtime marker set.
• Requires runtime/config/defaults.json during build finalization and release packaging.
• Produces actionable diagnostics when shipped defaults are missing while preserving safe embedded settings defaults.

Windows build stability

• Disables Nuitka link-time optimization for the official release build.
• Enables Nuitka low-memory mode and one compiler job.
• Preserves the locked Python 3.12, Nuitka 4.1.3, MSVC, OneDir, Playwright, and CustomTkinter architecture.
• Prevents the observed MSVC C1002: compiler is out of heap space in pass 2 and downstream LNK1257 peak-memory failure path.

Validation evidence

• 85 targeted export, browser-workflow, parser, checkpoint, archive, build, and runtime tests passed on Windows.
• 109 complete canonical tests passed on Windows.
• Python compilation, JSON/TOML parsing, whitespace checks, dependency locking, public-tree privacy verification, and isolated release-candidate verification passed.

The GitHub tag-triggered Windows build remains authoritative for the downloadable ContextVault-Windows-x64.zip.

ContextVault v0.2.0 — Export Reliability and Stability Update

Release date: 2026-07-31
Type: Bug-fix and reliability release
Platform: Windows 10/11 x64
Archive schema: 1.0

ContextVault v0.2.0 focuses on dependable export of long, virtualized, image-heavy ChatGPT conversations.

The release preserves the public archive layout while strengthening browser readiness, message checkpointing, asset routing, Windows filesystem behavior, concurrency, naming, cancellation, and validation.

Recommended for all users

v0.1.0 users should upgrade.

The portable release does not require Python.

Major improvements

Large conversations can continue while progress is real

The legacy export path could stop after a fixed 900-second readiness window even when hundreds of messages were still being discovered and checkpointed.

v0.2.0 uses a meaningful-progress stall policy:

• new messages reset progress;
• committed checkpoint counts reset progress;
• scroll movement and recovery state count as progress;
• total elapsed time alone does not fail the export;
• a true no-progress stall still produces an explicit error.

Message checkpoints before scrolling

Every stable virtualized window is processed before ContextVault scrolls away:

extract
→ write atomic message JSON
→ save exact code bytes
→ read back
→ verify
→ commit
→ scroll

Completed checkpoints are retained through one bounded recovery reload.

Message-specific retry and degraded preservation

A single malformed message no longer automatically discards an otherwise usable large conversation.

ContextVault retries the failed message, preserves prior checkpoints, reloads once when required, resumes the scan, and preserves an exhausted message as an explicit degraded record when safe.

Warnings are visible in metadata, manifest, Markdown, export log, and validation log.

Stalled images no longer freeze scrolling forever

A terminal broken image or persistent spinner is not treated as pending forever.

Image-specific loading indicators are separated from blocking ChatGPT loaders.

Grace periods:

Fast    8 seconds
Normal 20 seconds
Safe   45 seconds
Auto   20 seconds

After the grace period, the stable message can be checkpointed and scanning can continue with a warning.

Correct asset routing

• Decorative favicons are filtered.
• Image resources remain image resources.
• Attachment-control fallback is used only for attachments.
• An image HTTP error does not trigger a long UI search for an attachment control.

Safe browser workflow exclusivity

One complete browser workflow owns the browser context.

Duplicate exports and conflicting operations are rejected before they can interleave.

This prevents conversation switching during capture, inconsistent message counts, duplicate archive publication, and shared Playwright state corruption.

Correct archive titles

The scanned ChatGPT sidebar title is the canonical title.

ContextVault does not use the first assistant heading as the archive name.

When multiple conversations share a title, a stable conversation-identity suffix preserves both.

Atomic archive publication

Final destination naming is resolved when the validated staged archive is published.

A concurrent name collision no longer destroys staging or fails after a long successful build.

Windows path and file handling

• Short same-directory temporary filenames reduce Windows path pressure.
• Atomic JSON replacement handles transient sharing denials.
• Partial, invalid, checkpoint, browser-profile, and temporary runtime paths are ignored by Git.

Cleaner cancellation diagnostics

Expected user cancellation is logged as cancellation, not as a false browser failure traceback.

Other fixes

• Exact UTF-8 code-byte validation preserves CRLF/LF.
• Stable zero-message shells receive one bounded recovery reload.
• Current data-message-id message containers are supported.
• Launch Chrome uses the managed profile when Browser Profile Root is blank or points to regular daily Chrome data.
• Automatic fallback to an unrelated CDP browser is removed.
• Capture warnings are ordered and deduplicated.
• Archive validator preserves explicit warnings without hiding real corruption.
• Release-candidate verification now uses an isolated .venv-release environment and exact locked dependencies, preventing shared Python package drift from producing misleading release checks.

Validation evidence

For the source commit preceding release preparation:

• GitHub Actions Windows Python 3.12 CI passed.
• Environment verification passed.
• 81 forensic tests passed.
• Repository JSON/TOML parse checks passed.
• Python AST checks passed.
• Dependency synchronization checks passed.
• Security path tests passed.
• Archive, checkpoint, browser-worker, image-readiness, concurrency, and cancellation regressions passed.

The release asset is considered complete only when the tag-triggered Build and Release workflow successfully installs locked dependencies, runs tests, compiles Nuitka OneDir with MSVC, verifies the distribution, packages the ZIP, writes the SHA-256 file, and uploads both release assets.

Upgrade

Read Upgrading.

Preserve personal runtime data before replacing an old portable folder.

Download verification

Read Release verification.

Known limitations

Read Known limitations.

Support

Read SUPPORT.md.

v0.1.0 — Initial Development Release

🚀 ContextVault v0.1.0 — Initial Development Release

Welcome to the first development release of ContextVault.

ContextVault is an open-source desktop application designed to help users collect, organize, manage, and export AI conversations into a structured, portable archive. Whether you want to export ChatGPT chats, preserve AI research, create long-term knowledge bases, or back up conversations from multiple AI platforms, ContextVault aims to provide a reliable and production-ready solution.

This release establishes the engineering foundation of the project and prepares the repository for future feature development.

---

✨ Highlights

• Initial project architecture
• Production-oriented repository structure
• Engineering specification framework
• Enterprise coding standards
• Security policy and contribution guidelines
• Release management workflow
• GitHub Actions foundation
• Testing infrastructure placeholders
• Build configuration
• Documentation framework
• Open-source project governance

---

📚 Planned Features

ContextVault is being developed to support capabilities such as:

• Export ChatGPT conversations
• ChatGPT chat backup
• ChatGPT conversation export
• AI chat export
• AI conversation archive
• AI prompt history management
• AI knowledge base creation
• Structured conversation storage
• Local offline archive
• Portable conversation database
• Browser automation using Playwright
• Multi-session conversation management
• Searchable conversation archive
• Project-based organization
• Import and export workflows
• Long-term AI conversation preservation

---

🛠 Current Status

This is an early development release.

The application is not yet feature-complete. This version focuses on establishing a stable architecture, engineering standards, and development workflow that will support future production releases.

Core application functionality will be introduced progressively in upcoming releases.

---

🎯 Project Goals

ContextVault is designed with the following objectives:

• Build a reliable AI conversation archive
• Enable secure and structured conversation management
• Preserve AI chat history for long-term use
• Support reproducible and deterministic builds
• Maintain clean, scalable, and well-documented code
• Deliver a production-quality desktop application
• Encourage community-driven open-source development

---

🔍 Keywords

This project is relevant for users searching for:

• ChatGPT Chat Export
• Export ChatGPT Conversations
• ChatGPT Backup Tool
• ChatGPT Conversation Export
• ChatGPT History Export
• ChatGPT Chat Archive
• AI Chat Export
• AI Conversation Export
• AI Conversation Backup
• AI Conversation Archive
• AI Prompt Manager
• AI Knowledge Management
• AI Research Archive
• AI Conversation Organizer
• Playwright Automation
• Python Desktop Application
• Open Source AI Tools

---

🤝 Contributing

Contributions, bug reports, feature requests, and discussions are welcome.

Please review the project's documentation before contributing:

• CONTRIBUTING.md
• SECURITY.md
• project/README.md
• project/INDEX.md

---

⭐ Support the Project

If you find ContextVault useful or are interested in AI conversation management, consider:

• Starring the repository
• Reporting issues
• Suggesting new features
• Contributing code or documentation
• Sharing the project with the community

---

📄 License

ContextVault is released under the MIT License.

---

Thank you for supporting ContextVault. This initial development release marks the beginning of a production-focused, open-source platform for ChatGPT chat export, AI conversation backup, AI chat archive, and long-term AI knowledge management.