AuthorAware works in two tracks. Basic requires nothing beyond a browser — no Terminal, no installs. Full adds AI Chat and local file backup through a few quick Terminal commands.
Basic · no Terminal required
1
Browser Only
Open the file and write. Keystroke timing and composition score recorded automatically. Zero setup.
Ready now
2
GitHub Backup
Every checkpoint commits to a private repo. Video hashes timestamped for Verified/Proctored tiers — done entirely in your browser.
5 min setup
Full · a few quick commands in a Terminal
3
GitHub Permanent
Token stored on your machine — never lost when browser cache is cleared.
2 min
4
AI Collaborator
AI chat inside the console — Claude, GPT-4o, Gemini, Sonar and more.
10 min
5
Local Backup
Projects saved as real files on your computer. Survives browser clearing.
+2 min
6
Video Backup
Video files stored to GitHub LFS or S3-compatible bucket (S3, Backblaze B2). Optional — hashes always go to GitHub free.
+5 min
Basicno terminal needed
1
Level one
Browser Only
What you get
Full provenance recording — every keystroke, edit, pause, and paste tracked
Composition authenticity score — keystroke timing analysis sealed into every checkpoint
Trust tiers — select Basic, Verified (screen capture), or Proctored (screen + webcam) in-app with no setup
Attested checkpoints sealed with SHA-256 hashes
Export your ledger as a signed JSON file at any time (Ctrl+Shift+E)
Multiple projects, project switcher
Projects live in browser storage — clearing browser data deletes them
No AI assistant, no automatic backups
There is nothing to install. Open author-aware-v1.html in any modern browser and start writing.
What you'll see — AuthorAware
AuthorAware digital provenance for writersUntitled Project ▾
session active
☁ Backup⚙ Setup📖 User Guide0 words · 0 chars⟡ AI Chat💾 Local BackupAttest Checkpoint
Begin writing here. Your session is recorded quietly in the background.
About the startup warning. If no proxy is running when you open AuthorAware, a warning appears reminding you that your projects are in browser storage only. Clearing your browser's cache or site data will permanently delete them. Click "Download backup now" in that dialog to save a copy to your computer before writing.
The storage button in the top bar shows 📁 Browser in amber when you are in browser-only mode. Click it at any time to download a backup or learn how to upgrade to local storage.
Keeping your work safe — manual project folder
Browser storage is temporary. The safest habit in browser-only mode is to treat each downloaded backup as your real file — and keep them organized in a folder you control.
1
Create a project folder on your computer
Make a folder anywhere — Desktop, Documents, a USB drive. Give it a name that matches your project. This is where your attestation files will live.
2
Download your backup after each session
Press Ctrl+Shift+E (or ⌘+Shift+E on Mac) to export your project as a .json file. Move that file into your project folder. Each export is a complete snapshot — draft text, ledger, and all attested checkpoints.
3
Reload by dragging the folder into the console
To restore your work — on a new machine, after clearing the browser, or just to load an older version — open author-aware-v1.html and drag your project folder directly into the window. AuthorAware will scan it, find all the .json backups inside, and load them as projects. You can also drag a single .json file to restore just one.
Your project folder is your provenance archive. Keep it somewhere safe and back it up like any important file — cloud storage, external drive, or both.
2
Level two
GitHub Backup
What you get
Full provenance recording — every keystroke, edit, pause, and paste tracked
Every attested checkpoint commits automatically to a private GitHub repository
Timestamped, permanent record that exists independently of your machine or browser
For Verified/Proctored tiers: video hashes committed to GitHub at every checkpoint — server-side timestamp, no extra cost
Verifiable provenance on a neutral third-party platform — shareable with publishers or prize committees
Full version history: every checkpoint is a git commit, browsable in chronological order
No local file storage, no AI chat (add Level 4 for those)
GitHub is a free service for storing and versioning files. AuthorAware uses it as a provenance ledger: every time you attest a checkpoint, it commits that checkpoint to a private repository under your account. You control the repository — you decide who sees it.
1
Create a free GitHub account and a private repository
Go to github.com and sign up for a free account if you don't have one. Then create a new repository — name it something like my-novel-provenance. Set it to Private.
In AuthorAware, click ☁ Backup in the top bar. Enter your GitHub username, repository name, and the token you just copied. Click Save. From that point, every attested checkpoint commits to the repository automatically.
With GitHub connected, every checkpoint is committed to a private repository under your account — timestamped by a neutral platform that neither you nor anyone else can retroactively alter. That record is the one you share with a publisher, prize committee, or anyone else who needs to verify your work.
⚠ Your token is stored in the browser
If you clear your browser cache, your Personal Access Token will be deleted. It's easy to get a new one: go to github.com → Settings → Developer settings → Fine-grained tokens, find your token, and click Regenerate. Then re-enter it here.
To solve this permanently so the token survives cache clears, continue with the Full setup below.
Fullshort commands in a Terminal
Start here — do this once, applies to all Full levels
1
Open a terminal
A terminal is a text window where you type commands directly to your computer.
Mac
Windows
Linux
Press ⌘ Space to open Spotlight. Type Terminal and press Enter.
Press Windows + X, then click Terminal or press I.
Press Ctrl + Alt + T, or search your application menu for Terminal.
2
Check that Python is installed
In the terminal, type this and press Enter:
Mac
Windows
Linux
python3 --version
Python 3.11.4
python --version
Python 3.11.4
python3 --version
Python 3.11.4
You need version 3.8 or later. If Python is not installed, download it from python.org/downloads — the installer handles everything.
3
Download hap-proxy.py
Save it somewhere easy to find — Desktop or Documents works well.
4
Run the proxy
First, navigate to the folder where you saved hap-proxy.py. If you saved it to Downloads, type this and press Enter:
Mac
Windows
Linux
cd ~/Downloads
Then run:
python3 hap-proxy.py
HAP Proxy running — http://localhost:7432
Press Ctrl+C to stop.
cd C:\Users\[your-name]\Downloads
Replace [your-name] with your Windows username, then run:
python hap-proxy.py
HAP Proxy running — http://localhost:7432
Press Ctrl+C to stop.
cd ~/Downloads
Then run:
python3 hap-proxy.py
HAP Proxy running — http://localhost:7432
Press Ctrl+C to stop.
Leave this terminal window open while you write. Pick a level below to continue.
3
Level three
GitHub — fire and forget
What you get — everything in Level 2, plus:
Personal Access Token stored on your machine — never in the browser, never lost on cache clear
GitHub backup works automatically every session with zero re-entry
Proxy reads the token from disk on every start — you never touch it again
One command stores your GitHub credentials in ~/.hap-proxy/config.json on your machine. The proxy reads them on every start. From this point forward, GitHub backup is silent and permanent.
1
Run the GitHub setup
Open a new terminal window (leave the proxy running) and run:
Mac
Windows
Linux
python3 hap-proxy.py --github-setup
python hap-proxy.py --github-setup
python3 hap-proxy.py --github-setup
A window opens. Enter your GitHub username, repository name, and Personal Access Token. Click Save. Your token is written to disk — you will never need to enter it again.
2
Verify in AuthorAware
The ☁ Backup button in the top bar will show your repository name. Every attested checkpoint commits automatically — no re-entry ever, even after clearing your browser cache.
Done. Your GitHub backup is now permanent. Clear your browser cache any time — the token lives on your machine, not in the browser.
4
Level four
AI Collaborator
What you get — everything in Levels 1–3, plus:
AI chat panel inside AuthorAware — Claude, GPT-4o, Gemini, Sonar and more
AI exchanges sealed into a separate provenance transcript (visible, logged, honest)
Handoff summaries between sessions — the AI remembers your project context
Cross-provider conversations — switch models mid-session, context carries over
1
Get an API key from at least one AI provider
Pick one to start. Anthropic Claude works best with the console's writing context. Each provider charges a small per-use fee; typical usage costs a few cents per session.
You only need one key to get started. You can add more later.
2
Add your AI keys
Open a new terminal window (leave the proxy running) and run:
Mac
Windows
Linux
python3 hap-proxy.py --ai-setup
python hap-proxy.py --ai-setup
python3 hap-proxy.py --ai-setup
A window opens. Click + Add Key, choose your provider, paste your API key, and click Save & Close Setup. Keys are stored locally in ~/.hap-proxy/config.json — they never leave your machine.
3
Open AI Chat
In AuthorAware click ⟡ AI Chat. The panel will show a checkmark and the providers you configured will appear in the model selector.
With the proxy running and AI keys configured, your writing sessions include a transparent AI collaboration log — every exchange sealed into a separate provenance transcript alongside your human keystrokes.
5
Level five
Local Backup
What you get — everything in Level 4, plus:
Projects saved as JSON files in a real folder on your computer — survives browser clearing
Projects load automatically when you open AuthorAware
Storage button turns green and shows the folder name
Automatic project sync between browser and disk on every save
Local Backup uses the same proxy from the preamble above. Run it once with a folder picker to choose where your files go — from then on, every save and checkpoint writes to disk automatically.
1
Run the proxy with local setup
Stop the proxy if it's running (Ctrl+C), then run:
A folder picker opens first — choose where your projects will be saved, then the proxy starts. Leave this terminal window open while you write.
2
Return to AuthorAware
Refresh the page (or re-open the file). AuthorAware detects the proxy automatically — the storage button turns green and shows your folder name.
Storage button — before and after
Proxy not running
💾 Local Backup
→
Proxy connected — projects saved to disk
📁 .hap-proxy/projects
Once the proxy is running with a local folder, every attest checkpoint and every autosave writes to disk automatically. You can open that folder in Finder or File Explorer and see your project files directly.
6
Level six · optional
Video Backup
What you get — on top of Levels 1–5:
Actual video files (screen + webcam) stored off-device for disaster recovery
GitHub LFS option — videos alongside your provenance repo, same PAT, no new credentials
S3-compatible option — works with Amazon S3 or Backblaze B2 (~$0.04–$1.40/year for a full year of sessions)
In-app configuration — no CLI required once the proxy is running
You don't need this level for a valid provenance claim. When GitHub backup is connected (Level 2+) and you use Verified or Proctored tier, a video hash is committed to GitHub at every checkpoint — server-side timestamped and free. The actual video files stay on your machine. Level 6 adds off-device storage so those files survive disk failure or a lost device.
Option A — GitHub LFS
GitHub offers Large File Storage (LFS) as a paid add-on to any plan. Upgrading gives you up to 50 GB of LFS storage at ~$5/month. Your existing Personal Access Token works unchanged — there is nothing new to configure on the proxy side.
1
Enable LFS on your GitHub account
Go to github.com/settings/billing → Storage and Bandwidth → upgrade to a paid data pack. One data pack adds 50 GB storage + 50 GB bandwidth per month.
2
Select LFS in AuthorAware
In AuthorAware, select Verified or Proctored tier. The Video Provenance panel appears in the sidebar. Under "Video file backup", select GitHub LFS. Done — no proxy command needed.
Option B — S3-compatible (Amazon S3 or Backblaze B2)
Any S3-compatible bucket works: Amazon S3 (~$1.40/year at 6 GB), Backblaze B2 (~$0.04/year), or any S3-compatible provider. You need three things: an endpoint URL, an access key ID, and a secret access key.
1
Create a bucket and get credentials
Backblaze B2: Create a bucket at backblaze.com. Under Account → Application Keys, create a new key with read/write access to that bucket. Note the Key ID, Application Key, and S3 endpoint (format: https://s3.us-west-004.backblazeb2.com).
Amazon S3: Create a bucket in your AWS console. Create an IAM user with PutObject permission on that bucket. Note the Access Key ID, Secret Access Key, and your bucket's endpoint URL.
2
Enter credentials in AuthorAware
In AuthorAware, select Verified or Proctored tier. In the Video Provenance panel, select S3-compatible. Three fields appear: endpoint URL, access key, and secret. Fill them in and click Save credentials. AuthorAware stores them in browser storage and sends them to the proxy if running.
3
Optional — store credentials permanently on disk
To avoid re-entering credentials if you clear your browser, run:
Mac
Windows
Linux
python3 hap-proxy.py --video-backup-setup
python hap-proxy.py --video-backup-setup
python3 hap-proxy.py --video-backup-setup
A prompt walks you through choosing LFS, S3-compatible, or none. Credentials are saved to ~/.hap-proxy/config.json — never in the browser.
Backblaze B2 is the recommended option for most users: S3-compatible API, ~$0.004/GB/month, and free egress to Cloudflare-peered origins. At 6 GB/year of writing video, annual cost is roughly the price of one coffee.
Quick reference
Proxy commands
python3 hap-proxy.py
Start the proxy
python3 hap-proxy.py --github-setup
Store GitHub token on disk (fire and forget)
python3 hap-proxy.py --ai-setup
Add / update AI provider keys
python3 hap-proxy.py --local-setup
Choose storage folder
python3 hap-proxy.py --manage
List and delete projects
python3 hap-proxy.py --video-backup-setup
Configure LFS / S3 / Backblaze B2 video backup
Keyboard shortcuts
Ctrl+Shift+E Export project backup
Downloads full backup .json (draft + ledger)
Drag & drop Restore from backup
Drop a .json backup file onto the console
🐍 What is Python?
Python is a free, general-purpose programming language that runs directly on your computer. Think of it as a small engine that can power tools — the same way a browser runs web pages, Python runs scripts. hap-proxy.py is one such script: when you run it, Python starts a tiny private web server on your machine that AuthorAware communicates with. Nothing is sent to the internet. Nothing is installed permanently. Close the terminal window and it's gone.
Python comes pre-installed on most Macs and Linux systems. On Windows you may need to download it once from python.org/downloads — the installer handles everything, and it takes about two minutes.
Running Python code locally is what gives you genuine agency in provenance cases. You don't need any service provider, cloud account, or third party to prove the truth. Your proof lives on your machine — in a folder you control, signed by your own keystrokes, independent of whether this website exists tomorrow.