Per-signer private links
Each signer gets a 4-word passphrase that derives their own encrypted room. A leaked link affects exactly one signer's slot — never the whole document.
DocSign
Sign a PDF together, without a server in the middle.
Drop a PDF, mark each signer's zones, and hand out per-signer links. Every signer connects directly to your tab, signs, and the finished document carries a tamper-evident audit log. No accounts, no uploads, nothing sitting in our database.
Live
Why this exists
Every e-signature platform wants you to upload your contract, your signers' email addresses, and your signature image to their database — and pay per envelope for the privilege. The contract you signed last year is sitting in someone else's S3 bucket, indexed by your account. If you cancel, the audit trail goes with them.
DocSign is the small alternative. Your PDF never leaves your tab. Each signer gets a private, end-to-end-encrypted peer link. When everyone has signed, the finished document carries the audit log inside it — yours forever, no subscription required.
How it works
You open /bay/docsign/, drop a PDF, set the number of signers, and drag signature zones onto each page. The app hands you one secure link per signer.
Each link is bound to a unique 4-word passphrase. The signer opens it, their browser derives an encryption key from the passphrase via PBKDF2 (600,000 iterations), and joins a private WebRTC room with your tab. The PDF streams to them encrypted; they draw or upload their signature, and the placement comes back the same way.
Your tab merges each submission into the canonical PDF as it arrives, but holds it locally. Other signers see only the unsigned original — they never see each other's signatures while signing is in flight. Once the last zone is signed, your tab embeds an audit log (signer names, timestamps, SHA-256 of the original PDF) and broadcasts the finished document to every signer at once.
Privacy in one paragraph
The PDF and every signature image move directly between browser tabs over WebRTC DataChannels, encrypted with AES-256-GCM keys derived from the per-signer passphrase. The signaling tracker (used only to find the other tab) never sees encrypted payloads. The only persistence is your local IndexedDB, so a refresh resumes your session — but no copy of the document, the signatures, or the audit log ever reaches a Powerful Matter server.
What's in it
Real signatures, not a SaaS subscription.
Draw or upload
Mouse, trackpad, touchscreen, or stylus — Pointer Events unify them. Or upload a transparent PNG, a JPEG, or an SVG you saved before.
Signature reuse
Save your signature once as PNG (universal) or SVG (sharp at any size, bit-exact round-trip). Re-upload it next time.
Live status dashboard
Who's connected, who has signed, what's left. Per-signer "show link" if someone needs it again.
Download at any time
Initiator gets the canonical (current) PDF; signers get the unsigned original until completion. Once signed, the completed PDF goes to everyone.
Built-in audit log
The finished PDF carries the SHA-256 of the original, the list of signers with timestamps, and an attached audit.json readable by any PDF tool.
FAQ
Things people ask before they sign in here.
What happens if I close the tab mid-flow?
The session is restored from your IndexedDB on the next visit — same PDF, same zones, same passphrases. Signers' links keep working as soon as your tab is back online. They don't need to do anything; the WebRTC tracker re-announces every 30 seconds and reconnects automatically.
Are these signatures legally binding?
DocSign produces visible-signature PDFs with an embedded audit log (signer identities, timestamps, original-document hash). That meets the bar for typical commercial agreements where the parties trust each other and a documented signing event is required. It is not a qualified electronic signature under eIDAS — that requires a regulated trust service provider, which is by design out of scope for a no-server tool.
What if a signer is on a phone?
The signing flow works on touch devices — the canvas covers most of the viewport so there's room for an actual arm-motion signature. Marking signature zones (the initiator step) is more comfortable on a tablet+ screen.
What if my PDF is sensitive?
It never leaves your browser. The signaling tracker (which both tabs use to find each other) sees only the room ID hash, not the document. The PDF and every signature image are AES-256-GCM-encrypted with a key derived from the per-signer passphrase before they cross the wire.
Can a signer share their link with someone else?
Yes — there's no identity binding beyond "you have the passphrase". The per-signer link is the credential. Hand out links the way you'd hand out a one-time meeting code. If someone forwards their link, the recipient signs in their slot, and the signer name in the audit log is whatever name you (the initiator) wrote when setting up the session.