A place for everything.
Find something you saved. Pick up an old idea.
Recently added
Groups
Save a web page
Search everything
Search titles, groups, and the full text of imported sources.
Find relevant excerpts
Search your archive here. For summaries and connections, ask ChatGPT to use MindMesh in your conversation. This app makes no paid AI requests.
Make sense of what you saved.
Projects, ideas and useful next steps, grounded in your sources.
Loading saved review progress…
Identical text links to an existing review; its title, location and original file are not independently reviewed. AI reviews happen in your ChatGPT conversations.Automatic inventory & larger review batches
Loading inventory…
Content matching preserves every source. Folders and backup families are automatic suggestions, not AI conclusions. Empty notes remain visible.
Explore reviewed projects
Continue reviewing with ChatGPT
“Use MindMesh's bulk review packets to review a large batch of unique unread content. Read complete sources, connect related material and save the results.”
ChatGPT can read multiple complete notes per packet and save up to 250 concise reviews per request. Long sources still need all their pages read. Identical imported text is read once; automatic grouping never counts as an AI review.
Updates from chat
Changes arrive after your server's next sync. Refresh to see the latest delivery.
Change history & undo
Undo restores the previous organization. A later edit to any item blocks undo of that batch.
Drop it here.
Files, tabs, notes, conversations. Keep what matters to you.
Files or a whole folder
Drop them here, or choose from your device. Original files stay on your device.
Includes subfolders. You can drop several folders together, or add more after an import.
Supported formats & ZIP import
Markdown, text, HTML, JSON, CSV, TSV, logs, YAML, and XML: up to 2 MB each. PDF and Word (.docx): up to 20 MB each. ZIP: up to 100 MB, 5,000 entries, and 250 MB unpacked.
PDF and Word text is indexed; original uploaded files are kept for download. Images, audio, video, and scanned PDFs without a text layer are not read. Hidden and application support files are skipped with a reason. Nested ZIPs must be uploaded separately.
Folder imports are snapshots, not automatic sync. Reimport to add new or changed files; earlier versions stay available. For a ChatGPT export, extract and select conversations.json using ChatGPT export above.
All your tabs, in one go.
Paste a whole list from your browser, or choose a bookmark HTML / JSON / text export. Up to 10,000 links per batch. Titles and links become searchable; page text is not downloaded.
Nothing is saved until you choose Save links. Reimporting the same URLs skips links already saved here.
Get tabs from Android
Chrome: open the tab switcher, use Select tabs, then Share tabs. If your selection menu offers Select all, use it. Copy the shared text and paste it above. Tab groups may need a separate pass.
Firefox: if your tab menu offers Share all tabs, copy that list. If it does not, Export Tabs URLs is listed for Firefox Android and can copy titles and URLs or save a file. It needs permission to read your browser tabs. Choose its plain-text or HTML export, then paste or choose that file here.
MindMesh cannot see other apps’ tabs automatically. Browser menus vary by version. Compare the preview count with your browser’s count; grouped, inactive and private tabs may need separate handling. Importing never closes your tabs.
File results
Transfers
Check your inventory, find a file, or review what stays out.
Connect a computer
Download and extract the launcher, then open MindMesh-Import-Inventory.cmd. It requires Python 3.10 or newer with Tkinter. Choose your candidate manifest and paste a connection code below. Nothing transfers until you select Start.
Unknown formats remain visible. Files become searchable after extraction; ChatGPT organization happens when you request it. Trash hides sources owned only by trashed candidates from search and organization. Shared and previously imported sources remain available.
What gets uploaded?
Known technical files, videos, and archives (ZIP, 7z, RAR and other compressed containers) are excluded from automatic upload. Extract wanted documents locally before adding them to the inventory. MindMesh does not currently extract video content. Images and audio are held because MindMesh has no image analysis or transcription. Mixed application and game-library paths are held for review. Other personal content and unknown formats elsewhere remain eligible. These rules use paths, not file contents; archives and ordinary documents can still contain private information. Review individual held files below. No files on your PC are changed.
Choose a run to see filtering progress. Existing catalogues are checked automatically.
Choose run by ID
Legacy hide / restore
| Location / technical status | Handling | Size | |
|---|---|---|---|
| Enter the inventory run ID shown by the launcher, then refresh. | |||
Keep the useful stuff.
See the space in use and remove unwanted uploaded copies.
Free up space
Delete uploaded originals that have not become searchable sources. Files backing searchable items and shared originals are protected. Your PC is not accessed or changed.
Preview includes up to 200,000 matching catalogue records. Nothing is deleted until you confirm the preview.What stays after deletion?
A small record of the original path and deletion stays in your inventory so the launcher will not upload the same candidate again. Deleted file bytes are not kept in Trash. Existing searchable text, chat organization, and files shared with other records are protected by this cleanup tool.
Photos, image scans, and audio are now held before automatic upload because MindMesh does not analyze them. Text documents, conversation exports, and supported source files can still be imported. No paid AI service is involved.
Worth another look
A random selection of sources older than two weeks.