Use this tab to create a permanent, verifiable proof that a specific person sent you an email on a given date — before a deadline, before a decision, before a dispute. The stamp is stored on Arweave and cannot be altered or deleted. Anyone can verify it independently.

Enter your email address and an app password, then specify the sender and a date range. Click Fetch emails to retrieve matching messages, select one, and stamp it.

What you can prove

The stamp records the sender's address, the date, and a cryptographic hash of the message. It proves that a specific sender sent this exact message before the moment of stamping. It does not prove the content is true — only that it exists and was sent.

Gmail: create an app password

Gmail requires an app-specific password for IMAP — your regular Gmail password will not work here.

1. Make sure 2-Step Verification is enabled on your Google account.
2. Go to myaccount.google.com → Security → App passwords.
3. Choose Mail as the app and give it any name.
4. Copy the 16-character password shown and paste it into the Password field above.

The app password gives read-only access to your email via IMAP. You can revoke it any time from the same page.

Finding the sender and date in Gmail

1. Open the message in Gmail.
2. The sender's address is shown next to the name in the From line — hover over the name to reveal the full address.
3. The date is displayed at the top right of the message. For the exact timestamp, click the three-dot menu → Show original.
4. Enter the sender address and a date range that covers that date, then click Fetch emails.

Use this tab to draw a new conclusion from two or more existing Leima stamps. For each stamp, upload the manifest.json and verdict.pdf — both appear in the download row below any verdict. The system validates each manifest, reads the full verdict content, and lets the AI reason across all of them to evaluate your new combined claim.

For example: three reference letters stamped separately each prove one thing. A bundle can then prove "this person has three verified references."

What you can prove

The bundle verdict is itself a new permanent stamp. It proves that on this date, a set of independently sealed verdicts was examined and found to support your combined claim. The original stamps remain unchanged on Arweave — the bundle is an additional layer of reasoning on top of them.

Where to find the manifest file

After any stamp is created, a download row appears below the verdict with buttons for ⬇ manifest.json and ⬇ verdict.pdf. Save both files and upload them here — one pair per stamp, 2–10 pairs total. The manifest contains the sealed claim and metadata; the verdict PDF contains the full AI analysis.

Writing a good claim

Describe what the documents together prove — a timeline, an agreement, or a sequence of events across multiple sources. For example:

• "These documents show that the contractor delivered the work before the payment deadline."
• "Taken together, these stamps prove that both parties agreed to the terms before the contract was signed."

Upload a photo or paste an image URL, then write a claim about what it shows. The AI analyses the visual content and stamps the result permanently on Arweave. If the image contains C2PA provenance data, the stamp also records whether the image is authentic and unmodified.

What you can prove

The stamp records a hash of the image and the AI's verdict on your claim. It proves that this specific image existed and that the claim was evaluated at this point in time. If the image has valid C2PA data, the stamp additionally confirms that it has not been altered since it was captured.

What is C2PA?

C2PA (Coalition for Content Provenance and Authenticity) is a standard that embeds a cryptographically signed certificate inside an image at the moment of capture. It records the device, the time, and any edits made afterwards. If the image is modified after capture, the signature becomes invalid.

When a C2PA certificate is present and valid, the verdict states that the image's origin is cryptographically established. When it is invalid, the verdict notes that the image was modified after capture. When there is no C2PA data, the AI analyses the content only — no origin claim is made.

Which images have C2PA?

C2PA is supported by a growing number of cameras and apps:

• Cameras: Sony α9 III and newer, Leica M11-P, Nikon Z9 (firmware update), selected Canon EOS R models
• Smartphones: Some Qualcomm Snapdragon 8 Gen 3 devices, and apps like Truepic that embed C2PA at capture
• AI generators: Adobe Firefly, Microsoft Designer, and others that sign generated images to disclose AI origin

Most ordinary smartphone photos and screenshots do not contain C2PA data — they can still be stamped, but without origin verification.

Use this tab when you want to stamp a factual claim without attaching any source document. The AI evaluates the claim using its own knowledge and records the verdict permanently on Arweave. No URL, file, or email is needed.

Type your claim in the field above and click Analyse. The stamp proves that this specific claim was evaluated at this exact moment in time.

What you can prove

The stamp records the claim text and the AI's verdict. It proves that on this date, the claim was evaluated and found to be supported or not supported by AI knowledge available at that moment. This is useful for establishing that a fact was considered established — or disputed — at a given point in time.

When to use it

• Stamping a factual assertion you want to record at a specific date
• Establishing that a scientific consensus or historical fact was recognised before a dispute arose
• Creating a timestamped record of your own statement without needing to share any document

Limitations

The AI's knowledge has a training cutoff — very recent events may not be reflected. For claims about recent news, documents, or communications, use the Web, Email, or other source tabs instead. The stamp records the AI's verdict, not an independent fact-check.

Use this tab to stamp any document you have in hand. Use the sub-menu to choose how to bring it in: upload a PDF, paste text directly, upload a .txt file, or fetch a PDF from a URL. The AI evaluates your claim against the content and stamps the result permanently on Arweave.

Due diligence example

A company shares confidential financial projections during due diligence. The buyer stamps the document with a claim such as "this document shows projected revenue above €10M for 2026" — the verdict characterises the content without reproducing the sensitive figures in full.

The stamp records a cryptographic hash of the exact document. If either party later disputes what the document said, anyone can verify by fingerprint whether a given file matches the stamped source. A substituted or altered document will not match — revealing any attempt to misrepresent what was shared.

What to bring in

• Contracts, term sheets, or NDA clauses
• Financial statements or projections shared in confidence
• A statement or promise made in writing
• An excerpt from a report, article, or chat thread

Shorter, focused excerpts give more precise verdicts than entire documents. Use redaction in your claim to keep sensitive details out of the public verdict.

Paste a public URL and write a claim about what the page says. The AI fetches the page content at the moment you click Analyse and evaluates your claim against it. The result — including a hash of the fetched content — is stamped permanently on Arweave.

What you can prove

The stamp proves that at this specific moment, the page at this URL contained the content that supports or contradicts your claim. It captures a snapshot — not a live link. If the page is later edited or taken down, the stamp still stands as evidence of what it said at the time of stamping.

Tips for good results

• Use a direct link to the specific article or page, not a homepage or search result
• Pages behind a login, paywall, or CAPTCHA cannot be fetched
• For news articles, government documents, or court filings, the URL itself adds credibility
• If the page is very long, focus your claim on a specific statement so the AI can evaluate it precisely

Enter a GitHub repository and optionally a commit hash, branch, or tag. The AI fetches the specified files and evaluates your claim about the code. The result is stamped permanently — proving that this exact code existed in this repository at this point in time.

What you can prove

The stamp records the repository, the commit reference, and the AI's verdict. It proves that specific code — a function, a file, an algorithm — existed in a public repository at a given commit. This is useful for establishing an invention timeline, verifying a security claim at release time, or proving authorship.

You can also use this tab to stamp a code quality assessment: ask the AI to evaluate whether the code follows security best practices, is free of a specific class of vulnerability, or implements what its documentation claims. The stamp records both the code and the verdict at that exact version — useful for audits, vendor evaluations, or release sign-offs.

How to fill in the fields

• Repository: enter as owner/repo or paste the full GitHub URL
• Commit / branch / tag: leave as main to use the current HEAD, or enter a specific commit SHA or tag name to pin an exact version
• File paths: one path per line, up to 20 files — e.g. src/auth.py. Leave empty to use the repository README.

Only public repositories are supported without a token. For private repos, generate a GitHub personal access token with repo read scope and enter it in the token field.