
How to Find and Read Your Augment Code Chat History: A Step-by-Step Guide
Where Augment Code stores its chat transcripts, how to export them, and the fastest way to read long Augment sessions on desktop or mobile.
How to Find and Read Your Augment Code Chat History: A Step-by-Step Guide
Augment Code is one of the better AI coding assistants for working inside a real codebase. The context engine is strong, the inline completions are fast, and the chat panel reads your repo instead of hallucinating file paths. The one thing it does not do well is give you a readable archive of what you and the agent said to each other. If you have ever closed a long Augment session and later tried to find a specific decision, a snippet of refactor reasoning, or the exact prompt that unblocked you, you know the problem. The chat panel is built for working, not for rereading. This guide covers where Augment stores its transcripts, how to pull them out, and how to make a long session readable later.
Where Augment Code stores chat history on your machine
Augment runs as a VS Code and JetBrains extension, and chat state lives in two places. The first is a local workspace cache for the current session. The second is a cloud sync layer tied to your account for cross-device continuity. Understanding both matters because the fastest export method depends on which one you are reading from. The local cache gives you raw data instantly. The cloud layer gives you a tidy archive but takes a few minutes to generate.
The local cache lives inside the editor's extension storage directory, and the exact path depends on your operating system. Augment does not expose this folder in the extension UI, so you will need to open it directly from a terminal or file manager. The folder is safe to inspect while the editor is running, since the extension writes atomically. Here are the paths worth bookmarking:
- macOS (VS Code):
~/Library/Application Support/Code/User/globalStorage/augment.vscode-augment/ - Windows (VS Code):
%APPDATA%\Code\User\globalStorage\augment.vscode-augment\ - Linux (VS Code):
~/.config/Code/User/globalStorage/augment.vscode-augment/ - macOS (JetBrains):
~/Library/Application Support/JetBrains/<IDE>/augment/
Inside that folder you will find chat state in SQLite or JSON form, depending on your extension version. Newer builds use a local SQLite index with chat turns stored as JSON blobs, and a quick sqlite3 chat.db ".tables" will show you what is there. The schema is not officially documented but the field names are self-explanatory. Expect a threads table and a messages table with JSON blobs holding the actual turn content.
How to export a single Augment session as Markdown
The cleanest path for one conversation is the built-in copy flow. Open the chat panel, scroll to the top of the thread, and use the three-dot menu to pick "Copy conversation." Augment copies the full thread as Markdown. Paste it into a .md file and you have a portable transcript ready to read anywhere. This is the fastest option when you only want the current session and not the full archive.
That flow works for most sessions, but it falls apart once you have hundreds of turns, code blocks with mixed languages, or inline diff suggestions. The clipboard tends to flatten the structure, strip metadata, and sometimes collapse adjacent code fences. For those cases, pull straight from the local store or use the cloud export. The extra two minutes saves you from losing structure you cannot reconstruct later. Keep the Markdown file in a project folder named after the task, not the date, so you can find it again in a week.
How to export every Augment thread in bulk
For a bulk export you have two options. The first is the Augment web dashboard. Sign in at app.augmentcode.com, open Settings, and look for the data export option. It generates a zip of your chat history as JSON, usually ready in a few minutes. The archive includes every thread tied to your account across every device you have signed in from. Download the zip, extract it somewhere stable, and treat it as your long-term archive.
The second option is scripting against the local SQLite file. Run sqlite3 chat.db "SELECT json_extract(data, '$.messages') FROM threads;" > augment-threads.json and you have a raw JSON dump. Convert each thread to Markdown with a short Python script that walks the message array, emits ## User and ## Assistant headings, and preserves code fences. Fifty lines of Python gives you an archive you can read anywhere. This path is better when you need a programmatic transform or want to filter sessions by date or project before converting.
How to make long Augment sessions readable later
Raw exports solve the storage problem but not the reading problem. A real Augment session for a refactor can run thousands of lines, with dozens of code blocks, file paths, and inline suggestions. VS Code's preview pane is fine for short notes but not for a two-hour transcript. The typography is too tight, there is no outline navigation, and code blocks lose their visual weight once they fill the viewport. You need a reader built for long-form Markdown instead of a quick preview tool.
This is where a dedicated Markdown reader earns its keep. Prism MD renders long AI transcripts with the typography of a book: generous line height, proper code blocks with syntax highlighting, collapsible headings for navigation, and a reading mode that strips the noise. Drop your exported Augment Markdown into Prism MD and you can revisit the session instead of grep-scrolling through it. A few patterns help once your transcripts live outside the editor. Rename files to the task, so augment-stripe-webhook-refactor.md finds itself a week later. Keep code blocks fenced with the language tag intact for syntax highlighting. Add a two-line header at the top with the goal and the outcome.
For related workflows see how to read Cursor chat history and how to read Claude Code chat history, which cover the same export-and-read loop for other agents. If you want your Augment archive on your phone during a commute, the quickest path is a plain folder in iCloud Drive, Google Drive, or Dropbox. Export once, drop the Markdown files into the synced folder, and open them from a mobile reader. Prism MD opens straight from these folders on iOS and Android without needing a separate import step, and once a file has been opened it caches locally for offline reading. For a deeper look at keeping an AI archive portable, see sync AI conversations across devices.
FAQ: Augment Code chat history
Does Augment Code have a native export button? The web dashboard at app.augmentcode.com has a data export option that produces a JSON archive of all your chat history. Inside the editor, the three-dot menu on any thread includes a "Copy conversation" action that copies the thread as Markdown. Between those two paths you can export a single session or your full archive without any third-party tooling. Both are read-only operations and will not delete your original threads.
Where is Augment chat history stored locally? Augment stores a local cache inside the editor's extension storage. On macOS that is ~/Library/Application Support/Code/User/globalStorage/augment.vscode-augment/ for VS Code and the equivalent JetBrains config path for IntelliJ or PyCharm. On Windows the path lives under %APPDATA%\Code\User\globalStorage\augment.vscode-augment\. The files are SQLite or JSON depending on the extension version, and both formats are safe to read while the editor is open.
Can I read Augment transcripts offline on mobile? Yes, with a small setup. Export your sessions to Markdown, drop the files in iCloud Drive or Google Drive, and open them in a Markdown reader like Prism MD. Once a file has been opened it is cached locally for offline reading. This works on long flights, subway commutes, and anywhere else your laptop is not practical.
Will exporting delete my Augment chats? No, export is a read-only operation. Your original threads stay inside Augment unless you explicitly delete them from the dashboard or editor. You can export the same session as many times as you want without affecting the source data. This also means your exports can drift out of sync if you later edit a thread, so re-export whenever a session changes.
Create a free Prism MD account and start reading your AI chat history the way it was meant to be read. Your Augment sessions deserve better than a cramped IDE preview pane, and five minutes of setup is all it takes. Export once, drop the Markdown into a synced folder, and open it in Prism MD on whichever device you happen to have in hand. The first time you reread a session that unblocked a hard refactor, you will wonder why you ever let the IDE eat those transcripts.
Related reading
Ready to read your own AI documents?
Open ChatGPT, Claude, Gemini, or any markdown file in the reader built for the way models write.
- ✓Renders code, math & Mermaid out of the box
- ✓Works offline once you've opened a doc
- ✓Free forever for personal reading


