Email Source Text Exporter

How to use Email Source Text Exporter

Recover stored plain text and inspect alternative HTML as literal source records. Part IDs preserve MIME nesting, original header fields remain available, and filenames are treated as metadata. The tool never opens a link, executes an attachment, authenticates a sender or turns source extraction into a claim that mail is safe to share.

1. Choose a small supported CRLF .eml or load the original generated fixture.
2. Choose JSON part evidence or all supported inline plain-text parts, then run.
3. Review alternative variants, attachment metadata and personal information before saving or sharing the separate export.

Good to know

• One nonempty file at most 1 MiB, 20,000 transport lines, 998 bytes per line, 32 MIME parts, 6 levels, 200 total headers and 32 KiB headers per part. This profile requires CRLF transport and no literal NUL; arbitrary mailbox formats and repaired line endings are unsupported.
• Headers use ASCII source and selected UTF-8/US-ASCII RFC2047 B/Q words. Text defaults to strict US-ASCII when charset is absent; bytes above 127 are refused, never mapped through a Windows-1252 alias. Explicit UTF-8 uses fatal decoding and preserves BOM.
• Leaf transfer encodings are 7bit, 8bit, canonical complete Base64 or explicit quoted-printable escapes. Decoded leaves are at most 256 KiB each/512 KiB aggregate. Nonzero Base64 padding bits, incomplete escapes, unsupported literals and raw trailing QP whitespace are refused.
• Only multipart/mixed, multipart/alternative and multipart/related nesting is supported. Complete exact boundary lines are required. Nested message/*, signed/encrypted multipart semantics, RFC2231 extended parameters are explicitly refused. Exported text accepts only UTF-8/US-ASCII; omitted attachment bodies are not interpreted as character text. This is not a complete email client or RFC conformance validator.
• JSON keeps text/plain and literal text/html separately; PLAIN-TEXT exports supported nonattachment plain parts with part separators. It refuses HTML-only messages instead of inventing a visual transcript. Attachment bodies, including text leaves and descendants of attached multiparts, are omitted from every export. Unknown dispositions are treated as attachments. Declared headers, filenames and byte counts remain metadata. Raw headers, authors, addresses, text and filenames may contain personal data; no anonymization, sender authentication or malware certification occurs.
• Processing is local. Output is limited to 1,000,000 text characters unless a stricter file limit is stated.